diff --git a/docs/source/en/model_doc/afmoe.md b/docs/source/en/model_doc/afmoe.md index 4297701742aa..ecd0c726326e 100644 --- a/docs/source/en/model_doc/afmoe.md +++ b/docs/source/en/model_doc/afmoe.md @@ -13,7 +13,7 @@ specific language governing permissions and limitations under the License. rendered properly in your Markdown viewer. --> -*This model was released on {release_date} and added to Hugging Face Transformers on 2025-11-18.* +*This model was released on {release_date} and added to Hugging Face Transformers on 2025-11-29.*
diff --git a/docs/source/en/model_doc/fast_vlm.md b/docs/source/en/model_doc/fast_vlm.md index 25cbe3bff126..82ad34a6a8be 100644 --- a/docs/source/en/model_doc/fast_vlm.md +++ b/docs/source/en/model_doc/fast_vlm.md @@ -14,7 +14,7 @@ rendered properly in your Markdown viewer. --> -*This model was released on 2025-05-06 and added to Hugging Face Transformers on 2025-10-07.* +*This model was released on 2025-05-06 and added to Hugging Face Transformers on 2025-12-02.* # FastVLM diff --git a/docs/source/en/model_doc/glm46v.md b/docs/source/en/model_doc/glm46v.md index 6e099d7aaef6..64666cea7588 100644 --- a/docs/source/en/model_doc/glm46v.md +++ b/docs/source/en/model_doc/glm46v.md @@ -1,3 +1,20 @@ + +*This model was released on {release_date} and added to Hugging Face Transformers on 2025-11-15.* + # GLM-4.6V ## Glm46VConfig diff --git a/src/transformers/cache_utils.py b/src/transformers/cache_utils.py index 28f40952f2cd..055e9c396c8f 100644 --- a/src/transformers/cache_utils.py +++ b/src/transformers/cache_utils.py @@ -258,14 +258,35 @@ class StaticLayer(CacheLayerMixin): Args: max_cache_len (`int`): Maximum number of tokens that can be stored, used for tensor preallocation. + max_batch_size(`int`, *optional*): + Maximum batch size that can be stored """ is_compileable = True is_sliding = False - def __init__(self, max_cache_len: int): + def __init__(self, max_cache_len: int, max_batch_size: int | None = None): super().__init__() self.max_cache_len = max_cache_len + self.max_batch_size = max_batch_size + + def init_full_cache(self, key_states: torch.Tensor): + """Init the full batch with max_batch_size ahead of time.""" + if self.max_batch_size is None: + self.max_batch_size = key_states.shape[0] + _, self.num_heads, _, self.head_dim = key_states.shape + self.dtype, self.device = key_states.dtype, key_states.device + + self.keys_full = torch.zeros( + (self.max_batch_size, self.num_heads, self.max_cache_len, self.head_dim), + dtype=self.dtype, + device=self.device, + ) + self.values_full = torch.zeros( + (self.max_batch_size, self.num_heads, self.max_cache_len, self.head_dim), + dtype=self.dtype, + device=self.device, + ) def lazy_initialization(self, key_states: torch.Tensor): """ @@ -281,28 +302,27 @@ def lazy_initialization(self, key_states: torch.Tensor): i.e. `mode="reduce-overhead"` is known to fail). But it will in general work correctly, and prefill should not be compiled anyway for performances! """ - self.max_batch_size, self.num_heads, _, self.head_dim = key_states.shape - self.dtype, self.device = key_states.dtype, key_states.device - - self.keys = torch.zeros( - (self.max_batch_size, self.num_heads, self.max_cache_len, self.head_dim), - dtype=self.dtype, - device=self.device, - ) - self.values = torch.zeros( - (self.max_batch_size, self.num_heads, self.max_cache_len, self.head_dim), - dtype=self.dtype, - device=self.device, - ) - # Note: `mark_static_address` is used to tag the cache as a fixed data pointer, preventing compiled graph - # breaks when updating the cache. However, it is not supported when tracing the graph, so we skip it in this case. - # As prefill should never be compiled, this is not an issue and it will still be run (except when users compile - # prefill explicitly, but this should be avoided!) - if not is_torchdynamo_compiling(): - torch._dynamo.mark_static_address(self.keys) - torch._dynamo.mark_static_address(self.values) + self.init_full_cache(key_states) + + self.keys_map = [None] * (self.max_batch_size + 1) + self.values_map = [None] * (self.max_batch_size + 1) + for batch_size in range(1, self.max_batch_size + 1): + self.keys_map[batch_size] = self.keys_full[:batch_size] + self.values_map[batch_size] = self.values_full[:batch_size] + # Note: `mark_static_address` is used to tag the cache as a fixed data pointer, preventing compiled graph + # breaks when updating the cache. However, it is not supported when tracing the graph, so we skip it in this case. + # As prefill should never be compiled, this is not an issue and it will still be run (except when users compile + # prefill explicitly, but this should be avoided!) + if not is_torchdynamo_compiling(): + torch._dynamo.mark_static_address(self.keys_map[batch_size]) + torch._dynamo.mark_static_address(self.values_map[batch_size]) self.is_initialized = True + batch_size = key_states.shape[0] + + # Slice to current batch size + self.keys = self.keys_map[batch_size] + self.values = self.values_map[batch_size] def update( self, @@ -332,16 +352,29 @@ def update( cache_position if cache_position is not None else torch.arange(key_states.shape[-2], device=self.device) ) - # Update the cache + batch_size = key_states.shape[0] + self.keys = self.keys_map[batch_size] + self.values = self.values_map[batch_size] + + self._current_batch_size = batch_size + try: + # Try index_copy_ first (faster when it works) self.keys.index_copy_(2, cache_position, key_states) self.values.index_copy_(2, cache_position, value_states) - except NotImplementedError: + except (NotImplementedError, RuntimeError): # Fallback for devices like MPS where index_copy_ might not be supported. self.keys[:, :, cache_position] = key_states self.values[:, :, cache_position] = value_states + return self.keys, self.values + def reset(self): + if self.is_initialized: + self.keys_full.zero_() + self.values_full.zero_() + self._current_batch_size = None + def get_mask_sizes(self, cache_position: torch.Tensor) -> tuple[int, int]: """Return the length and offset of the cache, used to generate the attention mask""" kv_offset = 0 @@ -350,14 +383,28 @@ def get_mask_sizes(self, cache_position: torch.Tensor) -> tuple[int, int]: def get_seq_length(self) -> int: """Returns the sequence length of the cached states.""" - # Occupied cache == any slot in the 3rd dim (sequence length) holds a non-zero value. To save on compute, let's - # limit the check to the first batch member and head dimension. return (self.keys[0, 0].any(dim=-1)).sum() if self.is_initialized else 0 def get_max_cache_shape(self) -> int: """Return the maximum cache shape of the cache""" return self.max_cache_len + def reorder_cache(self, beam_idx: torch.LongTensor): + """Reorders this layer's cache for beam search.""" + if beam_idx.device != self.device: + beam_idx = beam_idx.to(self.device) + + current_batch_size = beam_idx.shape[0] + # Use the map to get the view for the current batch size + current_keys = self.keys_map[current_batch_size] + current_values = self.values_map[current_batch_size] + + new_keys = current_keys.index_select(0, beam_idx) + new_values = current_values.index_select(0, beam_idx) + + current_keys.copy_(new_keys) + current_values.copy_(new_values) + class StaticSlidingWindowLayer(StaticLayer): """ @@ -370,13 +417,15 @@ class StaticSlidingWindowLayer(StaticLayer): Maximum number of tokens that can be stored, used for tensor preallocation. sliding_window (`int`): The size of the sliding window. + max_batch_size(`int`, *optional*): + Maximum batch size that can be stored """ is_sliding = True - def __init__(self, max_cache_len: int, sliding_window: int): + def __init__(self, max_cache_len: int, sliding_window: int, max_batch_size: int | None = None): effective_max_cache_len = min(sliding_window, max_cache_len) - super().__init__(max_cache_len=effective_max_cache_len) + super().__init__(max_cache_len=effective_max_cache_len, max_batch_size=max_batch_size) self.cumulative_length = 0 def update( @@ -407,6 +456,10 @@ def update( cache_position if cache_position is not None else torch.arange(key_states.shape[-2], device=self.device) ) + batch_size = key_states.shape[0] + self.keys = self.keys_map[batch_size] + self.values = self.values_map[batch_size] + cumulative_length = self.cumulative_length is_full = cumulative_length >= self.max_cache_len # Update it now that we saved the value above @@ -425,10 +478,10 @@ def update( new_keys[:, :, index] = key_states new_values[:, :, index] = value_states - # Copy back into `self` (do not just assign again) in order to keep the static dynamo address - self.keys.copy_(new_keys) - self.values.copy_(new_values) - # Very important to return the `self` tensors here, as they have the static dynamo address + # Copy back into the batch-sliced portion (NOT the full cache) + self.keys[:batch_size].copy_(new_keys) + self.values[:batch_size].copy_(new_values) + # Return the batch-sliced view return self.keys, self.values # Already full but using more than 1 new token (e.g. prefill caching, chat continuation, etc...) else: @@ -451,13 +504,13 @@ def update( self.keys[:, :, cache_position] = key_states self.values[:, :, cache_position] = value_states - # Very important to return the `self` tensors here, as they have the static dynamo address + # Return the batch-sliced view return self.keys, self.values # We only cache the last `sliding_window` tokens - self.keys.copy_(full_key_states[:, :, -self.max_cache_len :, :]) - self.values.copy_(full_value_states[:, :, -self.max_cache_len :, :]) - # we should return the whole states instead of `self.keys/values` here, as otherwise we lose some context + self.keys[:batch_size].copy_(full_key_states[:, :, -self.max_cache_len :, :]) + self.values[:batch_size].copy_(full_value_states[:, :, -self.max_cache_len :, :]) + # we should return the whole states instead of sliced cache here, as otherwise we lose some context return full_key_states, full_value_states def get_mask_sizes(self, cache_position: torch.Tensor) -> tuple[int, int]: @@ -483,6 +536,12 @@ def get_seq_length(self) -> int: """Returns the sequence length of the cached states.""" return self.cumulative_length + def reset(self): + if self.is_initialized: + self.keys.zero_() + self.values.zero_() + self.cumulative_length = 0 + class QuantizedLayer(DynamicLayer): """ @@ -1019,6 +1078,8 @@ class StaticCache(Cache): offload_only_non_sliding (`bool`, *optional*, defaults to `True`): If `offloading` is `True`, this further decides if only the non-sliding layers will be offloaded (because usually the sliding layers are small in size, so there is no need to offload them, and skipping it is faster). + max_batch_size (`int`, *optional*): + The maximum batch size that will be used with this Cache . Example: @@ -1047,6 +1108,7 @@ def __init__( max_cache_len: int, offloading: bool = False, offload_only_non_sliding: bool = True, + max_batch_size: int | None = None, **kwargs, ): config = config.get_text_config(decoder=True) @@ -1066,19 +1128,39 @@ def __init__( layers = [] for layer_type in layer_types: if layer_type == "sliding_attention": - layer = StaticSlidingWindowLayer(max_cache_len=max_cache_len, sliding_window=config.sliding_window) + layer = StaticSlidingWindowLayer( + max_cache_len=max_cache_len, sliding_window=config.sliding_window, max_batch_size=max_batch_size + ) elif layer_type == "chunked_attention": # From a cache point of view, both sliding and chunked are the same in how they should behave and how many # states they should return - only the mask changes to make them different at the end! layer = StaticSlidingWindowLayer( - max_cache_len=max_cache_len, sliding_window=config.attention_chunk_size + max_cache_len=max_cache_len, + sliding_window=config.attention_chunk_size, + max_batch_size=max_batch_size, ) else: - layer = StaticLayer(max_cache_len=max_cache_len) + layer = StaticLayer(max_cache_len=max_cache_len, max_batch_size=max_batch_size) layers.append(layer) super().__init__(layers=layers, offloading=offloading, offload_only_non_sliding=offload_only_non_sliding) + def update( + self, + key_states: torch.Tensor, + value_states: torch.Tensor, + layer_idx: int, + cache_kwargs: Optional[dict[str, Any]] = None, + ) -> tuple[torch.Tensor, torch.Tensor]: + return self.layers[layer_idx].update(key_states, value_states, cache_kwargs) + + def reset(self): + for layer in self.layers: + layer.reset() + + def __len__(self): + return len(self.layers) + class QuantizedCache(Cache): """ diff --git a/tests_output.txt b/tests_output.txt new file mode 100644 index 000000000000..a23ed6ba3ab4 --- /dev/null +++ b/tests_output.txt @@ -0,0 +1,695 @@ +============================= test session starts ============================== +platform linux -- Python 3.12.3, pytest-8.4.2, pluggy-1.6.0 -- /home/vedth/stuhdy/transformers/venv/bin/python3 +cachedir: .pytest_cache +hypothesis profile 'default' +rootdir: /home/vedth/stuhdy/transformers +configfile: pyproject.toml +plugins: timeout-2.4.0, rich-0.2.0, xdist-3.8.0, rerunfailures-15.1, anyio-4.11.0, order-1.3.0, asyncio-1.3.0, hypothesis-6.148.3 +asyncio: mode=Mode.STRICT, debug=False, asyncio_default_fixture_loop_scope=function, asyncio_default_test_loop_scope=function +created: 2/2 workers +2 workers [191 items] + +scheduling tests via LoadScheduling + +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_load_with_global_device_set <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_fa2_generate <- tests/generation/test_utils.py +[gw1] [ 0%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_fa2_generate <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_fa3_generate <- tests/generation/test_utils.py +[gw1] [ 1%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_fa3_generate <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_00_fp16_pad_left_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 1%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_00_fp16_pad_left_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_01_fp16_pad_left <- tests/test_modeling_common.py +[gw1] [ 2%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_01_fp16_pad_left <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_02_fp16_pad_left_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 2%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_02_fp16_pad_left_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_03_fp16_pad_left_no_attn_mask <- tests/test_modeling_common.py +[gw1] [ 3%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_03_fp16_pad_left_no_attn_mask <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_04_fp16_pad_right_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 3%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_04_fp16_pad_right_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_05_fp16_pad_right <- tests/test_modeling_common.py +[gw1] [ 4%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_05_fp16_pad_right <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_06_fp16_pad_right_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 4%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_06_fp16_pad_right_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_07_fp16_pad_right_no_attn_mask <- tests/test_modeling_common.py +[gw1] [ 5%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_07_fp16_pad_right_no_attn_mask <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_08_fp32_pad_left_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 5%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_08_fp32_pad_left_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_09_fp32_pad_left <- tests/test_modeling_common.py +[gw1] [ 6%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_09_fp32_pad_left <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_10_fp32_pad_left_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 6%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_10_fp32_pad_left_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_11_fp32_pad_left_no_attn_mask <- tests/test_modeling_common.py +[gw1] [ 7%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_11_fp32_pad_left_no_attn_mask <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_12_fp32_pad_right_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 7%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_12_fp32_pad_right_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_13_fp32_pad_right <- tests/test_modeling_common.py +[gw1] [ 8%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_13_fp32_pad_right <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_14_fp32_pad_right_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 8%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_14_fp32_pad_right_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_15_fp32_pad_right_no_attn_mask <- tests/test_modeling_common.py +[gw1] [ 9%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_15_fp32_pad_right_no_attn_mask <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_16_bf16_pad_left_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 9%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_16_bf16_pad_left_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_17_bf16_pad_left <- tests/test_modeling_common.py +[gw1] [ 10%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_17_bf16_pad_left <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_18_bf16_pad_left_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 10%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_18_bf16_pad_left_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_19_bf16_pad_left_no_attn_mask <- tests/test_modeling_common.py +[gw1] [ 11%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_19_bf16_pad_left_no_attn_mask <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_20_bf16_pad_right_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 12%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_20_bf16_pad_right_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_21_bf16_pad_right <- tests/test_modeling_common.py +[gw1] [ 12%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_21_bf16_pad_right <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_22_bf16_pad_right_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +[gw1] [ 13%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_22_bf16_pad_right_no_attn_mask_sdpa_kernels <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_23_bf16_pad_right_no_attn_mask <- tests/test_modeling_common.py +[gw1] [ 13%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_23_bf16_pad_right_no_attn_mask <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_24_fp32_pad_left_output_attentions <- tests/test_modeling_common.py +[gw1] [ 14%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_matches_sdpa_inference_24_fp32_pad_left_output_attentions <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_padding_matches_padding_free_with_position_ids +[gw1] [ 14%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_eager_padding_matches_padding_free_with_position_ids +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_feed_forward_chunking <- tests/test_modeling_common.py +[gw1] [ 15%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_feed_forward_chunking <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attention_2_continue_generate_with_position_ids <- tests/generation/test_utils.py +[gw1] [ 15%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attention_2_continue_generate_with_position_ids <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids <- tests/generation/test_utils.py +[gw1] [ 16%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids_and_fa_kwargs <- tests/generation/test_utils.py +[gw1] [ 16%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attention_2_padding_matches_padding_free_with_position_ids_and_fa_kwargs <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attention_3_padding_matches_padding_free_with_position_ids <- tests/generation/test_utils.py +[gw1] [ 17%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attention_3_padding_matches_padding_free_with_position_ids <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attention_3_padding_matches_padding_free_with_position_ids_and_fa_kwargs <- tests/generation/test_utils.py +[gw1] [ 17%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attention_3_padding_matches_padding_free_with_position_ids_and_fa_kwargs <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break <- tests/test_modeling_common.py +[gw1] [ 18%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_can_compile_with_attention_mask_None_without_graph_break <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_can_dispatch_composite_models <- tests/test_modeling_common.py +[gw1] [ 18%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_can_dispatch_composite_models <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_equivalence <- tests/causal_lm_tester.py +[gw1] [ 19%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_equivalence <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_fp32_ln <- tests/test_modeling_common.py +[gw1] [ 19%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_fp32_ln <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_from_config <- tests/test_modeling_common.py +[gw1] [ 20%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_from_config <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_inference_equivalence <- tests/test_modeling_common.py +[gw1] [ 20%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_inference_equivalence <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_inference_equivalence_right_padding <- tests/test_modeling_common.py +[gw1] [ 21%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_2_inference_equivalence_right_padding <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_3_can_dispatch_composite_models <- tests/test_modeling_common.py +[gw1] [ 21%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_3_can_dispatch_composite_models <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_3_from_config <- tests/test_modeling_common.py +[gw1] [ 22%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_3_from_config <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_3_inference_equivalence <- tests/test_modeling_common.py +[gw1] [ 23%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_3_inference_equivalence <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_3_inference_equivalence_right_padding <- tests/test_modeling_common.py +[gw1] [ 23%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_3_inference_equivalence_right_padding <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_kernels_inference_equivalence <- tests/test_modeling_common.py +[gw1] [ 24%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_kernels_inference_equivalence <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_kernels_mps_inference_equivalence <- tests/test_modeling_common.py +[gw1] [ 24%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flash_attn_kernels_mps_inference_equivalence <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flex_attention_with_grads <- tests/test_modeling_common.py +[gw0] [ 25%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_load_with_global_device_set <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_attention_outputs <- tests/test_modeling_common.py +[gw0] [ 25%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_attention_outputs <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_attn_implementation_composite_models <- tests/test_modeling_common.py +[gw0] [ 26%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_attn_implementation_composite_models <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_batching_equivalence <- tests/test_modeling_common.py +[gw0] [ 26%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_batching_equivalence <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_bc_torch_dtype <- tests/test_modeling_common.py +[gw0] [ 27%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_bc_torch_dtype <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_be_initialized_on_meta <- tests/test_modeling_common.py +[gw0] [ 27%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_be_initialized_on_meta <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_init_all_missing_weights <- tests/test_modeling_common.py +[gw0] [ 28%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_init_all_missing_weights <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_load_from_already_mapped_keys <- tests/test_modeling_common.py +[gw0] [ 28%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_load_from_already_mapped_keys <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_load_ignoring_mismatched_shapes <- tests/test_modeling_common.py +[gw0] [ 29%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_load_ignoring_mismatched_shapes <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_load_with_device_context_manager <- tests/test_modeling_common.py +[gw0] [ 29%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_load_with_device_context_manager <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_set_attention_dynamically <- tests/test_modeling_common.py +[gw0] [ 30%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_set_attention_dynamically <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_set_attention_dynamically_composite_model <- tests/test_modeling_common.py +[gw0] [ 30%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_set_attention_dynamically_composite_model <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_use_safetensors <- tests/test_modeling_common.py +[gw0] [ 31%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_can_use_safetensors <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_cannot_load_with_meta_device_context_manager <- tests/test_modeling_common.py +[gw0] [ 31%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_cannot_load_with_meta_device_context_manager <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_causal_lm_can_accept_training_kwargs <- tests/causal_lm_tester.py +[gw0] [ 32%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_causal_lm_can_accept_training_kwargs <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_config <- tests/causal_lm_tester.py +[gw0] [ 32%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_config <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_config_attn_implementation_setter <- tests/test_modeling_common.py +[gw0] [ 33%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_config_attn_implementation_setter <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_correct_missing_keys <- tests/test_modeling_common.py +[gw0] [ 34%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_correct_missing_keys <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_cpu_offload <- tests/test_modeling_common.py +[gw0] [ 34%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_cpu_offload <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_custom_4d_attention_mask <- tests/generation/test_utils.py +[gw0] [ 35%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_custom_4d_attention_mask <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_determinism <- tests/test_modeling_common.py +[gw0] [ 35%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_determinism <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_disk_offload_bin <- tests/test_modeling_common.py +[gw0] [ 36%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_disk_offload_bin <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_disk_offload_safetensors <- tests/test_modeling_common.py +[gw0] [ 36%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_disk_offload_safetensors <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_rope_scaling_from_config_2_yarn <- tests/causal_lm_tester.py +[gw0] [ 37%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_rope_scaling_from_config_2_yarn <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_weights_reload_no_missing_tied_weights <- tests/test_modeling_common.py +[gw0] [ 37%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_weights_reload_no_missing_tied_weights <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_multi_gpu_data_parallel_forward <- tests/test_modeling_common.py +[gw0] [ 38%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_multi_gpu_data_parallel_forward <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_num_layers_is_small <- tests/test_modeling_common.py +[gw0] [ 38%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_num_layers_is_small <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_peft_gradient_checkpointing_enable_disable <- tests/test_modeling_common.py +[gw0] [ 39%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_peft_gradient_checkpointing_enable_disable <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_any_to_any <- tests/test_pipeline_mixin.py +[gw0] [ 39%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_any_to_any <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_any_to_any_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 40%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_any_to_any_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_audio_classification <- tests/test_pipeline_mixin.py +[gw0] [ 40%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_audio_classification <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_audio_classification_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 41%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_audio_classification_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_automatic_speech_recognition <- tests/test_pipeline_mixin.py +[gw0] [ 41%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_automatic_speech_recognition <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_automatic_speech_recognition_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 42%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_automatic_speech_recognition_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_depth_estimation <- tests/test_pipeline_mixin.py +[gw0] [ 42%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_depth_estimation <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_depth_estimation_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 43%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_depth_estimation_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_document_question_answering <- tests/test_pipeline_mixin.py +[gw0] [ 43%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_document_question_answering <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_document_question_answering_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 44%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_document_question_answering_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_feature_extraction <- tests/test_pipeline_mixin.py +[gw0] [ 45%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_feature_extraction <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_feature_extraction_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 45%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_feature_extraction_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_fill_mask <- tests/test_pipeline_mixin.py +[gw0] [ 46%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_fill_mask <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_fill_mask_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 46%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_fill_mask_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_classification <- tests/test_pipeline_mixin.py +[gw0] [ 47%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_classification <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_classification_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 47%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_classification_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_feature_extraction <- tests/test_pipeline_mixin.py +[gw0] [ 48%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_feature_extraction <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_feature_extraction_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 48%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_feature_extraction_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_segmentation <- tests/test_pipeline_mixin.py +[gw0] [ 49%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_segmentation <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_segmentation_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 49%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_segmentation_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_text_to_text <- tests/test_pipeline_mixin.py +[gw0] [ 50%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_text_to_text <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_text_to_text_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 50%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_text_to_text_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_to_text <- tests/test_pipeline_mixin.py +[gw0] [ 51%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_to_text <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_to_text_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 51%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_image_to_text_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_mask_generation <- tests/test_pipeline_mixin.py +[gw0] [ 52%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_mask_generation <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_mask_generation_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 52%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_mask_generation_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_object_detection <- tests/test_pipeline_mixin.py +[gw0] [ 53%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_object_detection <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_object_detection_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 53%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_object_detection_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_question_answering <- tests/test_pipeline_mixin.py +[gw0] [ 54%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_question_answering <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_question_answering_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 54%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_question_answering_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_summarization <- tests/test_pipeline_mixin.py +[gw0] [ 55%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_summarization <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_summarization_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 56%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_summarization_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_table_question_answering <- tests/test_pipeline_mixin.py +[gw0] [ 56%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_table_question_answering <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_table_question_answering_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 57%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_table_question_answering_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text2text_generation <- tests/test_pipeline_mixin.py +[gw0] [ 57%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text2text_generation <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text2text_generation_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 58%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text2text_generation_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_classification <- tests/test_pipeline_mixin.py +[gw0] [ 58%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_classification <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_classification_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 59%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_classification_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_generation <- tests/test_pipeline_mixin.py +[gw0] [ 59%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_generation <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_generation_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 60%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_generation_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_to_audio <- tests/test_pipeline_mixin.py +[gw0] [ 60%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_to_audio <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_to_audio_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 61%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_text_to_audio_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_token_classification <- tests/test_pipeline_mixin.py +[gw0] [ 61%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_token_classification <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_token_classification_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 62%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_token_classification_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_translation <- tests/test_pipeline_mixin.py +[gw0] [ 62%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_translation <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_translation_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 63%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_translation_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_video_classification <- tests/test_pipeline_mixin.py +[gw0] [ 63%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_video_classification <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_video_classification_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 64%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_video_classification_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_visual_question_answering <- tests/test_pipeline_mixin.py +[gw0] [ 64%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_visual_question_answering <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_visual_question_answering_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 65%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_visual_question_answering_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot <- tests/test_pipeline_mixin.py +[gw0] [ 65%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_audio_classification <- tests/test_pipeline_mixin.py +[gw0] [ 66%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_audio_classification <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_audio_classification_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 67%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_audio_classification_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 67%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_image_classification <- tests/test_pipeline_mixin.py +[gw0] [ 68%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_image_classification <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_image_classification_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 68%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_image_classification_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_object_detection <- tests/test_pipeline_mixin.py +[gw0] [ 69%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_object_detection <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_object_detection_fp16 <- tests/test_pipeline_mixin.py +[gw0] [ 69%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_pipeline_zero_shot_object_detection_fp16 <- tests/test_pipeline_mixin.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_problem_types <- tests/test_modeling_common.py +[gw0] [ 70%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_problem_types <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_question_answering_model <- tests/causal_lm_tester.py +[gw0] [ 70%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_question_answering_model <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_embeddings_untied <- tests/test_modeling_common.py +[gw0] [ 71%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_embeddings_untied <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_embeddings_untied_with_deepspeed <- tests/test_modeling_common.py +[gw0] [ 71%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_embeddings_untied_with_deepspeed <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_embeddings_untied_with_deepspeed_multi_gpu <- tests/test_modeling_common.py +[gw0] [ 72%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_embeddings_untied_with_deepspeed_multi_gpu <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_position_vector_embeddings <- tests/test_modeling_common.py +[gw0] [ 72%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_position_vector_embeddings <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_tokens_embeddings <- tests/test_modeling_common.py +[gw0] [ 73%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_tokens_embeddings <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_tokens_embeddings_with_deepspeed <- tests/test_modeling_common.py +[gw0] [ 73%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_tokens_embeddings_with_deepspeed <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_tokens_embeddings_with_deepspeed_multi_gpu <- tests/test_modeling_common.py +[gw0] [ 74%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_resize_tokens_embeddings_with_deepspeed_multi_gpu <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_retain_grad_hidden_states_attentions <- tests/test_modeling_common.py +[gw0] [ 74%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_retain_grad_hidden_states_attentions <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_reverse_loading_mapping <- tests/test_modeling_common.py +[gw0] [ 75%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_reverse_loading_mapping <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_save_load <- tests/test_modeling_common.py +[gw0] [ 75%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_save_load <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_save_load_keys_to_ignore_on_save <- tests/test_modeling_common.py +[gw0] [ 76%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_save_load_keys_to_ignore_on_save <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sdpa_can_compile_dynamic <- tests/test_modeling_common.py +[gw0] [ 76%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sdpa_can_compile_dynamic <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sdpa_can_dispatch_composite_models <- tests/test_modeling_common.py +[gw0] [ 77%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sdpa_can_dispatch_composite_models <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sdpa_can_dispatch_non_composite_models <- tests/test_modeling_common.py +[gw0] [ 78%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sdpa_can_dispatch_non_composite_models <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sdpa_can_dispatch_on_flash <- tests/test_modeling_common.py +[gw0] [ 78%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sdpa_can_dispatch_on_flash <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sdpa_padding_matches_padding_free_with_position_ids +[gw0] [ 79%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sdpa_padding_matches_padding_free_with_position_ids +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sequence_classification_model <- tests/causal_lm_tester.py +[gw0] [ 79%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sequence_classification_model <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sequence_classification_model_for_multi_label <- tests/causal_lm_tester.py +[gw0] [ 80%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sequence_classification_model_for_multi_label <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sequence_classification_model_for_single_label <- tests/causal_lm_tester.py +[gw0] [ 80%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sequence_classification_model_for_single_label <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sliding_window_mask <- tests/test_modeling_common.py +[gw0] [ 81%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_sliding_window_mask <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_tied_weights_keys <- tests/test_modeling_common.py +[gw0] [ 81%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_tied_weights_keys <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_token_classification_model <- tests/causal_lm_tester.py +[gw0] [ 82%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_token_classification_model <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_torch_compile_for_training <- tests/test_modeling_common.py +[gw0] [ 82%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_torch_compile_for_training <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_torch_export <- tests/test_modeling_common.py +[gw0] [ 83%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_torch_export <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_torch_save_load <- tests/test_modeling_common.py +[gw0] [ 83%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_torch_save_load <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_tp_plan_matches_params <- tests/test_modeling_common.py +[gw0] [ 84%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_tp_plan_matches_params <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_training <- tests/test_modeling_common.py +[gw0] [ 84%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_training <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_training_gradient_checkpointing <- tests/test_modeling_common.py +[gw0] [ 85%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_training_gradient_checkpointing <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_training_gradient_checkpointing_use_reentrant <- tests/test_modeling_common.py +[gw0] [ 85%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_training_gradient_checkpointing_use_reentrant <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_training_gradient_checkpointing_use_reentrant_false <- tests/test_modeling_common.py +[gw0] [ 86%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_training_gradient_checkpointing_use_reentrant_false <- tests/test_modeling_common.py +[gw1] [ 86%] FAILED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flex_attention_with_grads <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_forward_with_logits_to_keep <- tests/generation/test_utils.py +[gw1] [ 87%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_forward_with_logits_to_keep <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_from_pretrained_no_checkpoint <- tests/test_modeling_common.py +[gw1] [ 87%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_from_pretrained_no_checkpoint <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_generate_from_inputs_embeds <- venv/lib/python3.12/site-packages/_pytest/mark/structures.py +[gw1] [ 88%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_generate_from_inputs_embeds <- venv/lib/python3.12/site-packages/_pytest/mark/structures.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_generate_from_inputs_embeds_0_greedy <- tests/generation/test_utils.py +[gw1] [ 89%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_generate_from_inputs_embeds_0_greedy <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_generate_from_inputs_embeds_1_beam_search <- tests/generation/test_utils.py +[gw1] [ 89%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_generate_from_inputs_embeds_1_beam_search <- tests/generation/test_utils.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_generation_tester_mixin_inheritance <- tests/test_modeling_common.py +[gw1] [ 90%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_generation_tester_mixin_inheritance <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_gradient_checkpointing_backward_compatibility <- tests/test_modeling_common.py +[gw1] [ 90%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_gradient_checkpointing_backward_compatibility <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_gradient_checkpointing_enable_disable <- tests/test_modeling_common.py +[gw1] [ 91%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_gradient_checkpointing_enable_disable <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_hidden_states_output <- tests/test_modeling_common.py +[gw1] [ 91%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_hidden_states_output <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_inputs_embeds <- tests/test_modeling_common.py +[gw1] [ 92%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_inputs_embeds <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_inputs_embeds_matches_input_ids <- tests/test_modeling_common.py +[gw1] [ 92%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_inputs_embeds_matches_input_ids <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_internal_model_config_and_subconfig_are_same <- tests/test_modeling_common.py +[gw1] [ 93%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_internal_model_config_and_subconfig_are_same <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_keep_in_fp32_modules <- tests/test_modeling_common.py +[gw1] [ 93%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_keep_in_fp32_modules <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_load_save_without_tied_weights <- tests/test_modeling_common.py +[gw1] [ 94%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_load_save_without_tied_weights <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_load_with_mismatched_shapes <- tests/test_modeling_common.py +[gw1] [ 94%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_load_with_mismatched_shapes <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model <- tests/causal_lm_tester.py +[gw1] [ 95%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_base_model_prefix <- tests/test_modeling_common.py +[gw1] [ 95%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_base_model_prefix <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_get_set_embeddings <- tests/test_modeling_common.py +[gw1] [ 96%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_get_set_embeddings <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_is_small <- tests/test_modeling_common.py +[gw1] [ 96%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_is_small <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_main_input_name <- tests/test_modeling_common.py +[gw1] [ 97%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_main_input_name <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_outputs_equivalence +[gw1] [ 97%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_outputs_equivalence +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_parallelism <- tests/test_modeling_common.py +[gw1] [ 98%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_parallelism <- tests/test_modeling_common.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_rope_scaling_frequencies +[gw1] [ 98%] SKIPPED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_rope_scaling_frequencies +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_rope_scaling_from_config_0_linear <- tests/causal_lm_tester.py +[gw1] [ 99%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_rope_scaling_from_config_0_linear <- tests/causal_lm_tester.py +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_rope_scaling_from_config_1_dynamic <- tests/causal_lm_tester.py +[gw1] [100%] PASSED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_model_rope_scaling_from_config_1_dynamic <- tests/causal_lm_tester.py + +=================================== FAILURES =================================== +________________ AfmoeModelTest.test_flex_attention_with_grads _________________ +[gw1] linux -- Python 3.12.3 /home/vedth/stuhdy/transformers/venv/bin/python3 + +self = + + @require_torch_accelerator + def test_flex_attention_with_grads(self): + for model_class in self.all_model_classes: + config, inputs_dict = self.model_tester.prepare_config_and_inputs_for_common() + inputs_dict = self._prepare_for_class(inputs_dict, model_class) + model = model_class(config).to(device=torch_device) + + # If not all sub-models support flex, skip the test + if not all( + submodel._supports_flex_attn for submodel in model.modules() if isinstance(submodel, PreTrainedModel) + ): + self.skipTest(reason="At least some parts of this model do not support flex attention") + + # Set default attention to flex and update config values + config = self._prepare_config_headdim(config, 16) # specific to triton + + if model_class._can_set_attn_implementation(): + model = model_class(config).to(device=torch_device) + model.set_attn_implementation("flex_attention") + self.assertTrue(model.config._attn_implementation == "flex_attention") + else: + config._attn_implementation = "flex_attention" + model = model_class(config).to(device=torch_device) + + # Elaborate workaround for encoder-decoder models as some do not specify their main input + dummy_inputs = {model.main_input_name: inputs_dict[model.main_input_name].to(torch_device)} + for key in getattr(self, "additional_model_inputs", []): + # Some models don't have all `additional_model_inputs`, especially when we + # craft cases to test model in different settings + if key in inputs_dict: + dummy_inputs[key] = inputs_dict[key].to(torch_device) + + if config.is_encoder_decoder: + dummy_inputs["decoder_input_ids"] = inputs_dict["decoder_input_ids"].to(torch_device) + dummy_inputs["decoder_attention_mask"] = inputs_dict["decoder_attention_mask"].to(torch_device) + + # If this does not raise an error, the test passes (see https://github.com/huggingface/transformers/pull/35605) +> _ = model(**dummy_inputs) + ^^^^^^^^^^^^^^^^^^^^^ + +tests/test_modeling_common.py:3761: +_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ +venv/lib/python3.12/site-packages/torch/nn/modules/module.py:1775: in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/nn/modules/module.py:1786: in _call_impl + return forward_call(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +src/transformers/utils/generic.py:919: in wrapper + outputs = func(self, *args, **kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^ +src/transformers/models/afmoe/modeling_afmoe.py:629: in forward + hidden_states = decoder_layer( +src/transformers/modeling_layers.py:94: in __call__ + return super().__call__(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/nn/modules/module.py:1775: in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/nn/modules/module.py:1786: in _call_impl + return forward_call(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +src/transformers/models/afmoe/modeling_afmoe.py:479: in forward + hidden_states, _ = self.self_attn( +venv/lib/python3.12/site-packages/torch/nn/modules/module.py:1775: in _wrapped_call_impl + return self._call_impl(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/nn/modules/module.py:1786: in _call_impl + return forward_call(*args, **kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +src/transformers/models/afmoe/modeling_afmoe.py:415: in forward + output, attn_weights = attention_interface( +src/transformers/integrations/flex_attention.py:290: in flex_attention_forward + flex_attention_output = compile_friendly_flex_attention( +src/transformers/integrations/flex_attention.py:97: in compile_friendly_flex_attention + return flex_attention_compiled( +venv/lib/python3.12/site-packages/torch/_dynamo/eval_frame.py:845: in compile_wrapper + raise e.remove_dynamo_frames() from None # see TORCHDYNAMO_VERBOSE=1 + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/_inductor/compile_fx.py:990: in _compile_fx_inner + raise InductorError(e, currentframe()).with_traceback( +venv/lib/python3.12/site-packages/torch/_inductor/compile_fx.py:974: in _compile_fx_inner + mb_compiled_graph = fx_codegen_and_compile( +venv/lib/python3.12/site-packages/torch/_inductor/compile_fx.py:1695: in fx_codegen_and_compile + return scheme.codegen_and_compile(gm, example_inputs, inputs_to_check, graph_kwargs) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/_inductor/compile_fx.py:1505: in codegen_and_compile + compiled_module = graph.compile_to_module() + ^^^^^^^^^^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/_inductor/graph.py:2319: in compile_to_module + return self._compile_to_module() + ^^^^^^^^^^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/_inductor/graph.py:2329: in _compile_to_module + mod = self._compile_to_module_lines(wrapper_code) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/_inductor/graph.py:2397: in _compile_to_module_lines + mod = PyCodeCache.load_by_key_path( +venv/lib/python3.12/site-packages/torch/_inductor/codecache.py:3548: in load_by_key_path + mod = _reload_python_module(key, path, set_sys_modules=in_toplevel) + ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/_inductor/runtime/compile_tasks.py:33: in _reload_python_module + exec(code, mod.__dict__, mod.__dict__) +/tmp/torchinductor_vedth/ep/cep56kylllvniz3gr37e3fcgkrhmfv4xgxeohcmrpjftdphbcicz.py:624: in + async_compile.wait(globals()) +venv/lib/python3.12/site-packages/torch/_inductor/async_compile.py:631: in wait + self._wait_futures(scope) +venv/lib/python3.12/site-packages/torch/_inductor/async_compile.py:651: in _wait_futures + kernel = result.result() + ^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/_inductor/codecache.py:4289: in result + return self.result_fn() + ^^^^^^^^^^^^^^^^ +venv/lib/python3.12/site-packages/torch/_inductor/async_compile.py:470: in get_result + kernel.precompile( +venv/lib/python3.12/site-packages/torch/_inductor/runtime/triton_heuristics.py:451: in precompile + self._make_launchers() +_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ + +self = + + def _make_launchers(self): + if len(self.launchers) == len(self.compile_results): + return + + from torch._dynamo.device_interface import DeviceGuard + + device_interface = self.get_device_interface() + + # load binary to the correct device + with DeviceGuard(device_interface, self.triton_meta["device"]): + # need to initialize context + with dynamo_timed( + "CachingAutotuner.synchronize", + # Deliberately avoid overloading pt2_compile_events: + log_pt2_compile_event=False, + ): + device_interface.synchronize(device_interface.current_device()) + + launchers = [] + exc = None + for result in self.compile_results: + try: + launchers.append(result.make_launcher()) + + except (OutOfResources, PTXASError, torch.cuda.OutOfMemoryError) as e: + exc = e + if len(launchers) == 0: +> raise RuntimeError(f"No valid triton configs. {type(exc).__name__}: {exc}") +E torch._inductor.exc.InductorError: RuntimeError: No valid triton configs. OutOfMemoryError: out of resource: triton_tem_fused_0 Required: 147968 Hardware limit:101376 Reducing block sizes or `num_stages` may help. +E +E Set TORCHDYNAMO_VERBOSE=1 for the internal stack trace (please do this especially if you're reporting a bug to PyTorch). For even more developer context, set TORCH_LOGS="+dynamo" + +venv/lib/python3.12/site-packages/torch/_inductor/runtime/triton_heuristics.py:619: InductorError +=============================== warnings summary =============================== +:488 +:488 + :488: DeprecationWarning: builtin type SwigPyPacked has no __module__ attribute + +:488 +:488 + :488: DeprecationWarning: builtin type SwigPyObject has no __module__ attribute + +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_batching_equivalence + /home/vedth/stuhdy/transformers/tests/test_modeling_common.py:1091: UserWarning: Using a non-tuple sequence for multidimensional indexing is deprecated and will be changed in pytorch 2.9; use x[tuple(seq)] instead of x[seq]. In pytorch 2.9 this will be interpreted as tensor index, x[torch.tensor(seq)], which will result either in an error or a different result (Triggered internally at /pytorch/torch/csrc/autograd/python_variable_indexing.cpp:345.) + batched_row = batched_object[slice_ids] + +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flex_attention_with_grads +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flex_attention_with_grads + /home/vedth/stuhdy/transformers/venv/lib/python3.12/site-packages/torch/nn/attention/flex_attention.py:1083: DeprecationWarning: _compile flag on create_block_mask was originally added to work around a torch.compile limitation. That limitation has since been addressed. So, to compile create_block_mask, we suggest doing torch.compile(create_block_mask). This still works for now, but will be removed in the future. + warnings.warn( + +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flex_attention_with_grads + /home/vedth/stuhdy/transformers/venv/lib/python3.12/site-packages/torch/nn/attention/flex_attention.py:1622: FutureWarning: return_lse is deprecated and will be removed in v2.10. Please use return_aux=AuxRequest(lse=True) instead. + _warn_once( + +tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_generate_from_inputs_embeds + /usr/lib/python3.12/unittest/case.py:690: DeprecationWarning: It is deprecated to return a value that is not None from a test case (.decorator..wrapper of >) + return self.run(*args, **kwds) + +-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html +-- generated xml file: /home/vedth/stuhdy/transformers/test-results/junit.xml -- +=========================== short test summary info ============================ +SKIPPED [1] tests/generation/test_utils.py:1849: test is slow +SKIPPED [1] tests/generation/test_utils.py:1857: test requires Flash Attention 3 +SKIPPED [1] tests/models/afmoe/test_modeling_afmoe.py:107: Afmoe applies key/query norm which doesn't work with packing +SKIPPED [1] tests/generation/test_utils.py:1865: AfmoeForCausalLM does not support Flash Attention. +SKIPPED [1] tests/generation/test_utils.py:2082: test is slow +SKIPPED [1] tests/generation/test_utils.py:2089: test is slow +SKIPPED [1] tests/generation/test_utils.py:2098: test requires Flash Attention 3 +SKIPPED [1] tests/generation/test_utils.py:2105: test requires Flash Attention 3 +SKIPPED [1] tests/test_modeling_common.py:3367: test is slow +SKIPPED [1] tests/test_modeling_common.py:3294: This model is not a composite model! +SKIPPED [1] tests/causal_lm_tester.py:596: test is slow +SKIPPED [1] tests/test_modeling_common.py:3306: test requires bitsandbytes +SKIPPED [1] tests/test_modeling_common.py:3462: test is slow +SKIPPED [1] tests/test_modeling_common.py:2894: test is slow +SKIPPED [1] tests/test_modeling_common.py:2902: test is slow +SKIPPED [1] tests/test_modeling_common.py:3300: test requires Flash Attention 3 +SKIPPED [1] tests/test_modeling_common.py:3469: test requires Flash Attention 3 +SKIPPED [1] tests/test_modeling_common.py:2910: test requires Flash Attention 3 +SKIPPED [1] tests/test_modeling_common.py:2918: test requires Flash Attention 3 +SKIPPED [1] tests/test_modeling_common.py:2876: test is slow +SKIPPED [1] tests/test_modeling_common.py:2884: test requires MPS +SKIPPED [1] tests/test_modeling_common.py:2926: Model is not a composite model. +SKIPPED [1] tests/test_modeling_common.py:4162: No conversion found for this model +SKIPPED [1] tests/test_modeling_common.py:3981: This model is not composite +SKIPPED [1] tests/generation/test_utils.py:2149: AfmoeForCausalLM is not guaranteed to work with custom 4D attention masks +SKIPPED [1] tests/test_modeling_common.py:2327: test requires multiple CUDA GPUs +SKIPPED [1] tests/test_pipeline_mixin.py:595: AfmoeModelTest::test_pipeline_any_to_any_float32 is skipped: `any-to-any` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:601: AfmoeModelTest::test_pipeline_any_to_any_float16 is skipped: `any-to-any` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:493: AfmoeModelTest::test_pipeline_audio_classification_float32 is skipped: `audio-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:497: AfmoeModelTest::test_pipeline_audio_classification_float16 is skipped: `audio-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:502: AfmoeModelTest::test_pipeline_automatic_speech_recognition_float32 is skipped: `automatic-speech-recognition` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:506: AfmoeModelTest::test_pipeline_automatic_speech_recognition_float16 is skipped: `automatic-speech-recognition` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:511: AfmoeModelTest::test_pipeline_depth_estimation_float32 is skipped: `depth-estimation` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:518: AfmoeModelTest::test_pipeline_depth_estimation_float16 is skipped: `depth-estimation` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:525: test requires PyTesseract +SKIPPED [1] tests/test_pipeline_mixin.py:532: test requires PyTesseract +SKIPPED [1] tests/test_pipeline_mixin.py:539: AfmoeModelTest::test_pipeline_feature_extraction_float32 is skipped: Could not find any model architecture in the tiny models JSON file for `feature-extraction`. +SKIPPED [1] tests/test_pipeline_mixin.py:543: AfmoeModelTest::test_pipeline_feature_extraction_float16 is skipped: Could not find any model architecture in the tiny models JSON file for `feature-extraction`. +SKIPPED [1] tests/test_pipeline_mixin.py:548: AfmoeModelTest::test_pipeline_fill_mask_float32 is skipped: `fill-mask` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:552: AfmoeModelTest::test_pipeline_fill_mask_float16 is skipped: `fill-mask` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:557: AfmoeModelTest::test_pipeline_image_classification_float32 is skipped: `image-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:563: AfmoeModelTest::test_pipeline_image_classification_float16 is skipped: `image-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:618: AfmoeModelTest::test_pipeline_image_feature_extraction_float32 is skipped: `image-feature-extraction` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:625: AfmoeModelTest::test_pipeline_image_feature_extraction_float16 is skipped: `image-feature-extraction` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:569: AfmoeModelTest::test_pipeline_image_segmentation_float32 is skipped: `image-segmentation` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:576: AfmoeModelTest::test_pipeline_image_segmentation_float16 is skipped: `image-segmentation` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:583: AfmoeModelTest::test_pipeline_image_text_to_text_float32 is skipped: `image-text-to-text` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:589: AfmoeModelTest::test_pipeline_image_text_to_text_float16 is skipped: `image-text-to-text` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:607: AfmoeModelTest::test_pipeline_image_to_text_float32 is skipped: `image-to-text` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:612: AfmoeModelTest::test_pipeline_image_to_text_float16 is skipped: `image-to-text` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:632: `run_pipeline_test` is currently not implemented. +SKIPPED [1] tests/test_pipeline_mixin.py:639: `run_pipeline_test` is currently not implemented. +SKIPPED [1] tests/test_pipeline_mixin.py:646: AfmoeModelTest::test_pipeline_object_detection_float32 is skipped: `object-detection` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:653: AfmoeModelTest::test_pipeline_object_detection_float16 is skipped: `object-detection` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:660: AfmoeModelTest::test_pipeline_question_answering_float32 is skipped: `question-answering` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:664: AfmoeModelTest::test_pipeline_question_answering_float16 is skipped: `question-answering` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:669: AfmoeModelTest::test_pipeline_summarization_float32 is skipped: `summarization` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:673: AfmoeModelTest::test_pipeline_summarization_float16 is skipped: `summarization` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:678: AfmoeModelTest::test_pipeline_table_question_answering_float32 is skipped: `table-question-answering` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:682: AfmoeModelTest::test_pipeline_table_question_answering_float16 is skipped: `table-question-answering` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:687: AfmoeModelTest::test_pipeline_text2text_generation_float32 is skipped: `text2text-generation` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:691: AfmoeModelTest::test_pipeline_text2text_generation_float16 is skipped: `text2text-generation` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:696: AfmoeModelTest::test_pipeline_text_classification_float32 is skipped: `text-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:700: AfmoeModelTest::test_pipeline_text_classification_float16 is skipped: `text-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:705: AfmoeModelTest::test_pipeline_text_generation_float32 is skipped: Could not find any model architecture in the tiny models JSON file for `text-generation`. +SKIPPED [1] tests/test_pipeline_mixin.py:710: AfmoeModelTest::test_pipeline_text_generation_float16 is skipped: Could not find any model architecture in the tiny models JSON file for `text-generation`. +SKIPPED [1] tests/test_pipeline_mixin.py:715: AfmoeModelTest::test_pipeline_text_to_audio_float32 is skipped: `text-to-audio` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:720: AfmoeModelTest::test_pipeline_text_to_audio_float16 is skipped: `text-to-audio` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:725: AfmoeModelTest::test_pipeline_token_classification_float32 is skipped: `token-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:729: AfmoeModelTest::test_pipeline_token_classification_float16 is skipped: `token-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:734: AfmoeModelTest::test_pipeline_translation_float32 is skipped: `translation` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:738: AfmoeModelTest::test_pipeline_translation_float16 is skipped: `translation` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:743: AfmoeModelTest::test_pipeline_video_classification_float32 is skipped: `video-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:750: AfmoeModelTest::test_pipeline_video_classification_float16 is skipped: `video-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:757: AfmoeModelTest::test_pipeline_visual_question_answering_float32 is skipped: `visual-question-answering` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:763: AfmoeModelTest::test_pipeline_visual_question_answering_float16 is skipped: `visual-question-answering` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:769: AfmoeModelTest::test_pipeline_zero_shot_float32 is skipped: `zero-shot` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:778: AfmoeModelTest::test_pipeline_zero_shot_audio_classification_float32 is skipped: `zero-shot-audio-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:783: AfmoeModelTest::test_pipeline_zero_shot_audio_classification_float16 is skipped: `zero-shot-audio-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:773: AfmoeModelTest::test_pipeline_zero_shot_float16 is skipped: `zero-shot` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:788: AfmoeModelTest::test_pipeline_zero_shot_image_classification_float32 is skipped: `zero-shot-image-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:793: AfmoeModelTest::test_pipeline_zero_shot_image_classification_float16 is skipped: `zero-shot-image-classification` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:799: AfmoeModelTest::test_pipeline_zero_shot_object_detection_float32 is skipped: `zero-shot-object-detection` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/test_pipeline_mixin.py:805: AfmoeModelTest::test_pipeline_zero_shot_object_detection_float16 is skipped: `zero-shot-object-detection` is not in `self.pipeline_model_mapping` for `AfmoeModelTest`. +SKIPPED [1] tests/causal_lm_tester.py:404: Model does not support question answering +SKIPPED [1] tests/test_modeling_common.py:1908: test requires deepspeed +SKIPPED [1] tests/test_modeling_common.py:1920: test requires deepspeed +SKIPPED [1] tests/test_modeling_common.py:1561: Model does not have position embeddings +SKIPPED [1] tests/test_modeling_common.py:1796: test requires deepspeed +SKIPPED [1] tests/test_modeling_common.py:1808: test requires deepspeed +SKIPPED [1] tests/test_modeling_common.py:4084: No conversion found for this model +SKIPPED [1] tests/test_modeling_common.py:3167: test is slow +SKIPPED [1] tests/test_modeling_common.py:3023: AfmoeModel does not support SDPA +SKIPPED [1] tests/test_modeling_common.py:3086: test is slow +SKIPPED [1] tests/models/afmoe/test_modeling_afmoe.py:111: Afmoe applies key/query norm which doesn't work with packing +SKIPPED [1] tests/causal_lm_tester.py:341: Model does not support sequence classification +SKIPPED [1] tests/causal_lm_tester.py:370: Model does not support sequence classification +SKIPPED [1] tests/causal_lm_tester.py:355: Model does not support sequence classification +SKIPPED [1] tests/test_modeling_common.py:3476: Model does not support sliding window mask +SKIPPED [1] tests/causal_lm_tester.py:387: Model does not support token classification +SKIPPED [1] tests/test_modeling_common.py:3530: test is slow +SKIPPED [1] tests/test_modeling_common.py:3580: test is slow +SKIPPED [1] tests/test_modeling_common.py:4042: Model does not have a TP plan. +SKIPPED [1] tests/test_modeling_common.py:3896: No subconfigs so the test does not make sense +SKIPPED [1] tests/models/afmoe/test_modeling_afmoe.py:119: Afmoe has moe, output can be different +SKIPPED [1] tests/test_modeling_common.py:2492: test requires multiple accelerators +SKIPPED [1] tests/models/afmoe/test_modeling_afmoe.py:115: Afmoe applies key/query norm which doesn't work with packing +FAILED tests/models/afmoe/test_modeling_afmoe.py::AfmoeModelTest::test_flex_attention_with_grads - torch._inductor.exc.InductorError: RuntimeError: No valid triton configs. OutOfMemoryError: out of resource: triton_tem_fused_0 Required: 147968 Hardware limit:101376 Reducing block sizes or `num_stages` may help. + +Set TORCHDYNAMO_VERBOSE=1 for the internal stack trace (please do this especially if you're reporting a bug to PyTorch). For even more developer context, set TORCH_LOGS="+dynamo" +======= 1 failed, 83 passed, 107 skipped, 9 warnings in 73.49s (0:01:13) =======