Add llm generated explanations to `TAGDataset` #9918

xnuohz · 2025-01-05T10:32:47Z

Issue

Script

python examples/llm/glem.py --dataset arxiv --train_without_ext_pred --text_type llm_explanation

for more information, see https://pre-commit.ci

puririshi98 · 2025-01-06T22:55:45Z

this is awesome, but can you add an example to run? (also make sure the CI checks are all green)

…ch_geometric into tagdataset/add-llm-exp-pred

xnuohz · 2025-01-07T17:05:52Z

Thanks, @puririshi98 Can you clarify more about the runnable example? A simple way is to apply it to the GLEM model. However, a better way is to add a TAPE model, which is worth opening a new PR for easy review.
btw, there is a little issue with GLEM's example. Also plz take a look if you have bandwidth:)

Hi @akihironitta CI error is weird, seems like the error has nothing to do with my changes, can you help take a look?

puririshi98 · 2025-01-07T19:01:54Z

@xnuohz i think for now just adding it as an optional flag to GLEM example is okay. feel free to submit a seperate PR for TAPE. plz ping me on slack since i have github emails heavily filtered otherwise my inbox would explode. feel free to include this as a flag there as well when you do it

puririshi98 · 2025-01-07T19:03:20Z

Hi @akihironitta CI error is weird, seems like the error has nothing to do with my changes, can you help take a look?

@xnuohz ignore those for now, i was previously just talking about the linters that were red, your CI was functionally green before. once you address my above comment im sure these new issues will go away since their unrelated to your code, ive had this happen to me many times and they always go away on future respins. just my experience.

xnuohz · 2025-01-10T10:31:36Z

Namespace(gpu=0, num_runs=10, num_em_iters=1, dataset='arxiv', text_type='llm_explanation', pl_ratio=0.5, hf_model='prajjwal1/bert-tiny', gnn_model='SAGE', gnn_hidden_channels=256, gnn_num_layers=3, gat_heads=4, lm_batch_size=256, gnn_batch_size=1024, external_pred_path=None, alpha=0.5, beta=0.5, lm_epochs=10, gnn_epochs=50, gnn_lr=0.002, lm_lr=0.001, patience=3, verbose=False, em_order='lm', lm_use_lora=False, token_on_disk=False, out_dir='output/', train_without_ext_pred=True)
Running on: NVIDIA GeForce RTX 3090
/home/ubuntu/Softwares/anaconda3/envs/pyg-dev/lib/python3.9/site-packages/ogb/nodeproppred/dataset_pyg.py:69: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.
  self.data, self.slices = torch.load(self.processed_paths[0])
/home/ubuntu/Projects/pytorch_geometric/torch_geometric/data/in_memory_dataset.py:300: UserWarning: It is not recommended to directly access the internal storage format `data` of an 'InMemoryDataset'. If you are absolutely certain what you are doing, access the internal storage via `InMemoryDataset._data` instead to suppress this warning. Alternatively, you can access stacked individual attributes of every graph via `dataset.{attr_name}`.
  warnings.warn(msg)
Processing...
Done!
Tokenizing Text Attributed Graph raw_text: 100%|█████████████████████████████████████████████████████████████████| 169343/169343 [00:22<00:00, 7604.87it/s]
Tokenizing Text Attributed Graph llm_explanation: 100%|██████████████████████████████████████████████████████████| 169343/169343 [00:20<00:00, 8320.77it/s]
40 ['node-feat.csv.gz', 'node-label.csv.gz', 'ogbn-arxiv.csv', 'num-edge-list.csv.gz', 'num-node-list.csv.gz', 'node-gpt-response.csv.gz', 'edge.csv.gz', 'node_year.csv.gz', 'node-text.csv.gz']
train_idx: 136411, gold_idx: 90941, pseudo labels ratio: 0.5, 0.49999450192982264
Building language model dataloader...-->done
GPU memory usage -- data to gpu: 0.10 GB
build GNN dataloader(GraphSAGE NeighborLoader)--># GNN Params: 217640
2025-01-10 01:08:52.467527: I tensorflow/core/util/port.cc:111] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
2025-01-10 01:08:52.485008: E tensorflow/compiler/xla/stream_executor/cuda/cuda_dnn.cc:9342] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2025-01-10 01:08:52.485033: E tensorflow/compiler/xla/stream_executor/cuda/cuda_fft.cc:609] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2025-01-10 01:08:52.485046: E tensorflow/compiler/xla/stream_executor/cuda/cuda_blas.cc:1518] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
2025-01-10 01:08:52.488697: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
2025-01-10 01:08:52.887660: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
Some weights of BertForSequenceClassification were not initialized from the model checkpoint at prajjwal1/bert-tiny and are newly initialized: ['classifier.bias', 'classifier.weight']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
# LM Params: 4391080
pretraining gnn to generate pseudo labels
Epoch: 01 Loss: 2.1608 Approx. Train: 0.4124
Epoch: 02 Loss: 1.5093 Approx. Train: 0.5615
Epoch: 03 Loss: 1.3932 Approx. Train: 0.5870
Epoch: 04 Loss: 1.3258 Approx. Train: 0.6046
Epoch: 05 Loss: 1.2801 Approx. Train: 0.6159
Train: 0.6067, Val: 0.5852
Epoch: 06 Loss: 1.2459 Approx. Train: 0.6250
Train: 0.6145, Val: 0.5911
Epoch: 07 Loss: 1.2151 Approx. Train: 0.6317
Train: 0.6196, Val: 0.5999
Epoch: 08 Loss: 1.1907 Approx. Train: 0.6374
Train: 0.6213, Val: 0.5876
Epoch: 09 Loss: 1.1649 Approx. Train: 0.6445
Train: 0.6297, Val: 0.6033
Epoch: 10 Loss: 1.1433 Approx. Train: 0.6514
Train: 0.6290, Val: 0.5988
Epoch: 11 Loss: 1.1221 Approx. Train: 0.6560
Train: 0.6420, Val: 0.5989
Epoch: 12 Loss: 1.0989 Approx. Train: 0.6615
Train: 0.6392, Val: 0.6019
Pretrain Early stopped by Epoch: 12
Pretrain gnn time: 10.77s
Saved predictions to output/preds/arxiv/gnn_pretrain.pt
Pretraining acc: 0.6392, Val: 0.6019, Test: 0.5453
EM iteration: 1, EM phase: lm
Move lm model from cpu memory
Epoch 01 Loss: 1.5116 Approx. Train: 0.6574
Epoch 02 Loss: 1.1643 Approx. Train: 0.7199
Epoch 03 Loss: 1.0531 Approx. Train: 0.7243
Epoch 04 Loss: 0.9468 Approx. Train: 0.7283
Epoch 05 Loss: 0.8540 Approx. Train: 0.7320
Train: 0.8205, Val: 0.6925,
Epoch 06 Loss: 0.7706 Approx. Train: 0.7373
Train: 0.8343, Val: 0.6895,
Epoch 07 Loss: 0.7037 Approx. Train: 0.7413
Train: 0.8464, Val: 0.6699,
Epoch 08 Loss: 0.6463 Approx. Train: 0.7451
Train: 0.8590, Val: 0.6741,
Epoch 09 Loss: 0.6028 Approx. Train: 0.7487
Train: 0.8680, Val: 0.6777,
Early stopped by Epoch: 9,                             Best acc: 0.6925400181214135
EM iteration: 2, EM phase: gnn
Move gnn model from cpu memory
Epoch: 01 Loss: 0.9413 Approx. Train: 0.6264
Epoch: 02 Loss: 0.9080 Approx. Train: 0.6299
Epoch: 03 Loss: 0.8870 Approx. Train: 0.6345
Epoch: 04 Loss: 0.8745 Approx. Train: 0.6363
Epoch: 05 Loss: 0.8623 Approx. Train: 0.6394
Train: 0.6444, Val: 0.6100,
Epoch: 06 Loss: 0.8464 Approx. Train: 0.6423
Train: 0.6546, Val: 0.6163,
Epoch: 07 Loss: 0.8352 Approx. Train: 0.6439
Train: 0.6560, Val: 0.6143,
Epoch: 08 Loss: 0.8229 Approx. Train: 0.6460
Train: 0.6628, Val: 0.6180,
Epoch: 09 Loss: 0.8094 Approx. Train: 0.6495
Train: 0.6485, Val: 0.6083,
Epoch: 10 Loss: 0.7965 Approx. Train: 0.6522
Train: 0.6647, Val: 0.6136,
Epoch: 11 Loss: 0.7855 Approx. Train: 0.6547
Train: 0.6693, Val: 0.6173,
Early stopped by Epoch: 11,                             Best acc: 0.6179737575086413
Best GNN validation acc: 0.6179737575086413,LM validation acc: 0.6925400181214135
============================
Best test acc: 0.6018352776577578, model: lm
Total running time: 0.08 hours

puririshi98

Looks good but can you make an argparser option that combines the raw text and the llm explanation and run that? curious how it effects accuracy

xnuohz · 2025-07-19T13:31:56Z

@puririshi98 @akihironitta Inspector is broken in this CI, any thoughts?
inspector.type_repr(Optional[Tensor]) == 'Optional[typing.Any]'

torch_geometric/utils/_coalesce.py

test/test_inspector.py

puririshi98 · 2025-09-04T04:31:48Z

@xnuohz i recommend analysing the current latest CI and suggesting some potential solutions and i can review and see if we can help unblock you

xnuohz · 2025-09-04T14:13:57Z

@puririshi98 In my local env, found that running individual tests without errors, but running all tests simultaneously resulted in the same type inference errors as CI/CD, which suggests that PyTorch's JIT type inference system may be corrupted, causing Optional[Tensor] to be incorrectly inferred as Optional[Any]. I think still need to clean the cache. It is wrong to modify the assert result directly as in PR.

Cleaning the cache once in the workflow configuration is useless; consider cleaning it for each test file. So I added a JIT cache cleanup function to conftest, but it had no effect.
Optional[List[str]] is not type safe for JIT, removed it from TAGDataset params and it works. but text is also Optional[List[str]], I don't know why it didn't throw a CI error
PyG uses the Any type extensively, not sure if this is a good choice.

codecov · 2025-09-06T11:03:24Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 85.16%. Comparing base (c211214) to head (dfbe8a4).
⚠️ Report is 112 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #9918      +/-   ##
==========================================
- Coverage   86.11%   85.16%   -0.95%     
==========================================
  Files         496      510      +14     
  Lines       33655    35952    +2297     
==========================================
+ Hits        28981    30620    +1639     
- Misses       4674     5332     +658

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

…ch_geometric into tagdataset/add-llm-exp-pred

puririshi98 · 2025-09-09T15:28:46Z

can you rerun w/ and w/o tag dataset on using the latest nvidia container: https://catalog.ngc.nvidia.com/orgs/nvidia/containers/pyg/tags
want to make sure recent changes didnt break anything

xnuohz · 2025-09-10T16:43:21Z

root@7df2f109d384:/workspace/pytorch_geometric# python examples/llm/glem.py --dataset arxiv --train_without_ext_pred --text_type llm_explanation
  WARNING: This test may require more RAM than available.
    Estimated RAM needed: ~80 GB
    Detected available RAM: 53.78 GB
    If the program crashes or is killed, consider upgrading system memory.
Namespace(gpu=0, num_runs=10, num_em_iters=1, dataset='arxiv', text_type='llm_explanation', pl_ratio=0.5, hf_model='prajjwal1/bert-tiny', gnn_model='SAGE', gnn_hidden_channels=256, gnn_num_layers=3, gat_heads=4, lm_batch_size=256, gnn_batch_size=1024, external_pred_path=None, alpha=0.5, beta=0.5, lm_epochs=10, gnn_epochs=50, gnn_lr=0.002, lm_lr=0.001, patience=3, verbose=False, em_order='lm', lm_use_lora=False, token_on_disk=False, out_dir='output/', train_without_ext_pred=True)
Running on: NVIDIA GeForce RTX 3090
config.json: 100%|███████████████████████████████████████████████████████████████████████████████████████| 285/285 [00:00<00:00, 920kB/s]
vocab.txt: 232kB [00:00, 310kB/s] 
Processing...
Done!
Tokenizing Text Attributed Graph raw_text: 100%|███████████████████████████████████████████████| 169343/169343 [00:27<00:00, 6071.79it/s]
Tokenizing Text Attributed Graph llm_explanation: 100%|████████████████████████████████████████| 169343/169343 [00:25<00:00, 6523.98it/s]
Tokenizing Text Attributed Graph all: 100%|████████████████████████████████████████████████████| 169343/169343 [00:30<00:00, 5513.98it/s]
40 ['node-feat.csv.gz', 'node-label.csv.gz', 'ogbn-arxiv.csv', 'num-edge-list.csv.gz', 'num-node-list.csv.gz', 'node-gpt-response.csv.gz', 'edge.csv.gz', 'node_year.csv.gz', 'node-text.csv.gz']
train_idx: 136411, gold_idx: 90941, pseudo labels ratio: 0.5, 0.49999450192982264
Building language model dataloader...-->done
GPU memory usage -- data to gpu: 0.10 GB
build GNN dataloader(GraphSAGE NeighborLoader)--># GNN Params: 217640
pytorch_model.bin: 100%|████████████████████████████████████████████████████████████████████████████| 17.8M/17.8M [00:14<00:00, 1.24MB/s]
Some weights of BertForSequenceClassification were not initialized from the model checkpoint at prajjwal1/bert-tiny and are newly initialized: ['classifier.bias', 'classifier.weight']
You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.
# LM Params: 4391080
pretraining gnn to generate pseudo labels
Epoch: 01 Loss: 2.1597 Approx. Train: 0.4125
Epoch: 02 Loss: 1.5095 Approx. Train: 0.5615
Epoch: 03 Loss: 1.3933 Approx. Train: 0.5868
Epoch: 04 Loss: 1.3252 Approx. Train: 0.6044
model.safetensors:   0%|                                                                                     | 0.00/17.7M [00:00<?, ?B/s]Epoch: 05 Loss: 1.2792 Approx. Train: 0.6153
Train: 0.6063, Val: 0.5837
Epoch: 06 Loss: 1.2447 Approx. Train: 0.6254
Train: 0.6140, Val: 0.5914
Epoch: 07 Loss: 1.2134 Approx. Train: 0.6319
Train: 0.6178, Val: 0.5995
Epoch: 08 Loss: 1.1900 Approx. Train: 0.6371
Train: 0.6218, Val: 0.5906
Epoch: 09 Loss: 1.1641 Approx. Train: 0.6447
Train: 0.6299, Val: 0.6017
Epoch: 10 Loss: 1.1423 Approx. Train: 0.6520
Train: 0.6313, Val: 0.6020
Epoch: 11 Loss: 1.1204 Approx. Train: 0.6563
Train: 0.6418, Val: 0.5989
Epoch: 12 Loss: 1.0975 Approx. Train: 0.6620
Train: 0.6400, Val: 0.6032
Epoch: 13 Loss: 1.0795 Approx. Train: 0.6667
Train: 0.6495, Val: 0.6043
Epoch: 14 Loss: 1.0631 Approx. Train: 0.6719
Train: 0.6524, Val: 0.6068
Epoch: 15 Loss: 1.0440 Approx. Train: 0.6767
Train: 0.6584, Val: 0.6067
Epoch: 16 Loss: 1.0269 Approx. Train: 0.6801
Train: 0.6598, Val: 0.6040
Pretrain Early stopped by Epoch: 16
Pretrain gnn time: 11.01s
Saved predictions to output/preds/arxiv/gnn_pretrain.pt
Pretraining acc: 0.6598, Val: 0.6040, Test: 0.5534
EM iteration: 1, EM phase: lm
Move lm model from cpu memory
model.safetensors: 100%|████████████████████████████████████████████████████████████████████████████| 17.7M/17.7M [00:13<00:00, 1.34MB/s]
Epoch 01 Loss: 1.5224 Approx. Train: 0.6629
Epoch 02 Loss: 1.1747 Approx. Train: 0.7225
Epoch 03 Loss: 1.0573 Approx. Train: 0.7288
Epoch 04 Loss: 0.9507 Approx. Train: 0.7316
Epoch 05 Loss: 0.8558 Approx. Train: 0.7367
Train: 0.8136, Val: 0.6888,
Epoch 06 Loss: 0.7725 Approx. Train: 0.7413
Train: 0.8381, Val: 0.6975,
Epoch 07 Loss: 0.7026 Approx. Train: 0.7456
Train: 0.8511, Val: 0.6841,
Epoch 08 Loss: 0.6509 Approx. Train: 0.7483
Train: 0.8618, Val: 0.6741,
Epoch 09 Loss: 0.6005 Approx. Train: 0.7522
Train: 0.8742, Val: 0.6783,
Epoch 10 Loss: 0.5563 Approx. Train: 0.7548
Train: 0.8817, Val: 0.6764,
Early stopped by Epoch: 10,                             Best acc: 0.6974730695660928
EM iteration: 2, EM phase: gnn
Move gnn model from cpu memory
Epoch: 01 Loss: 0.8856 Approx. Train: 0.6402
Epoch: 02 Loss: 0.8523 Approx. Train: 0.6455
Epoch: 03 Loss: 0.8358 Approx. Train: 0.6488
Epoch: 04 Loss: 0.8235 Approx. Train: 0.6492
Epoch: 05 Loss: 0.8067 Approx. Train: 0.6531
Train: 0.6642, Val: 0.6129,
Epoch: 06 Loss: 0.7988 Approx. Train: 0.6534
Train: 0.6687, Val: 0.6145,
Epoch: 07 Loss: 0.7797 Approx. Train: 0.6562
Train: 0.6690, Val: 0.6125,
Epoch: 08 Loss: 0.7666 Approx. Train: 0.6603
Train: 0.6752, Val: 0.6095,
Epoch: 09 Loss: 0.7573 Approx. Train: 0.6615
Train: 0.6749, Val: 0.6076,
Epoch: 10 Loss: 0.7471 Approx. Train: 0.6629
Train: 0.6804, Val: 0.6150,
Epoch: 11 Loss: 0.7333 Approx. Train: 0.6665
Train: 0.6788, Val: 0.6128,
Early stopped by Epoch: 11,                             Best acc: 0.6150206382764523
Best GNN validation acc: 0.6150206382764523,LM validation acc: 0.6974730695660928
============================
Best test acc: 0.5963417896014649, model: lm
Total running time: 0.10 hours

puririshi98

lgtm

puririshi98 · 2025-09-10T16:58:00Z

lgtm @akihironitta @rusty1s @wsad1 to merge

xnuohz added 10 commits December 29, 2024 19:28

update

f330b06

update

6b1c556

Merge branch 'fix/glem-example' into tape

5caf0c3

add llm explanation token

f1c7931

add lm training

5c76117

update

a252312

Merge branch 'master' into tape

6da5717

update

e302bd4

update

82c9450

update

29f3881

xnuohz requested a review from wsad1 as a code owner January 5, 2025 10:32

xnuohz and others added 4 commits January 5, 2025 18:33

add changelog

b139ac2

[pre-commit.ci] auto fixes from pre-commit.com hooks

0f643d5

for more information, see https://pre-commit.ci

update

07fe2ec

[pre-commit.ci] auto fixes from pre-commit.com hooks

695ed01

for more information, see https://pre-commit.ci

xnuohz added 4 commits January 7, 2025 23:00

Merge branch 'master' into tagdataset/add-llm-exp-pred

9e65389

Merge branch 'tagdataset/add-llm-exp-pred' of github.com:xnuohz/pytor…

b064ab5

…ch_geometric into tagdataset/add-llm-exp-pred

update

74f2bb9

update

502659a

puririshi98 and others added 3 commits January 7, 2025 11:22

Merge branch 'master' into tagdataset/add-llm-exp-pred

d09912b

Merge branch 'master' into tagdataset/add-llm-exp-pred

1b3e679

update

ac402f3

Merge branch 'master' into tagdataset/add-llm-exp-pred

ac7bfed

puririshi98 requested changes Jan 11, 2025

View reviewed changes

update

fda21c2

puririshi98 requested changes Jul 22, 2025

View reviewed changes

torch_geometric/utils/_coalesce.py Outdated Show resolved Hide resolved

xnuohz commented Jul 22, 2025

View reviewed changes

test/test_inspector.py Outdated Show resolved Hide resolved

xnuohz and others added 7 commits July 22, 2025 17:55

Merge branch 'master' into tagdataset/add-llm-exp-pred

b993f9d

update

6aa63a8

update

6aeea47

update

15d8d5c

update

400e61f

update

4daee4e

Merge branch 'master' into tagdataset/add-llm-exp-pred

8feeea2

puririshi98 requested a review from rusty1s as a code owner August 25, 2025 18:14

puririshi98 added 3 commits August 26, 2025 19:18

Merge branch 'master' into tagdataset/add-llm-exp-pred

efb7af1

Merge branch 'master' into tagdataset/add-llm-exp-pred

5a9afc7

Merge branch 'master' into tagdataset/add-llm-exp-pred

34053c6

xnuohz added 2 commits September 6, 2025 18:52

fix ci

2acd83d

Merge branch 'master' into tagdataset/add-llm-exp-pred

c978b6c

xnuohz added 4 commits September 6, 2025 19:09

fix ci

46f6dcf

Merge branch 'tagdataset/add-llm-exp-pred' of github.com:xnuohz/pytor…

fef296e

…ch_geometric into tagdataset/add-llm-exp-pred

fix ci

74d5fed

Merge branch 'master' into tagdataset/add-llm-exp-pred

dfbe8a4

puririshi98 approved these changes Sep 10, 2025

View reviewed changes

akihironitta approved these changes Sep 19, 2025

View reviewed changes

puririshi98 merged commit f4bca53 into pyg-team:master Sep 19, 2025
18 checks passed

xnuohz deleted the tagdataset/add-llm-exp-pred branch September 20, 2025 03:51

Add llm generated explanations to TAGDataset #9918

Add llm generated explanations to TAGDataset #9918

Uh oh!

Conversation

xnuohz commented Jan 5, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Issue

Script

Uh oh!

puririshi98 commented Jan 6, 2025

Uh oh!

xnuohz commented Jan 7, 2025

Uh oh!

puririshi98 commented Jan 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

puririshi98 commented Jan 7, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

xnuohz commented Jan 10, 2025

Uh oh!

puririshi98 left a comment

Choose a reason for hiding this comment

Uh oh!

xnuohz commented Jul 19, 2025

Uh oh!

Uh oh!

Uh oh!

puririshi98 commented Sep 4, 2025

Uh oh!

xnuohz commented Sep 4, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov bot commented Sep 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

puririshi98 commented Sep 9, 2025

Uh oh!

xnuohz commented Sep 10, 2025

Uh oh!

puririshi98 left a comment

Choose a reason for hiding this comment

Uh oh!

puririshi98 commented Sep 10, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Add llm generated explanations to `TAGDataset` #9918

Add llm generated explanations to `TAGDataset` #9918

xnuohz commented Jan 5, 2025 •

edited

Loading

puririshi98 commented Jan 7, 2025 •

edited

Loading

puririshi98 commented Jan 7, 2025 •

edited

Loading

xnuohz commented Sep 4, 2025 •

edited

Loading

codecov bot commented Sep 6, 2025 •

edited

Loading