Skip to content

Fix MLX multi-batch validation memory growth #32

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
0hq merged 2 commits into from
Mar 18, 2026
+14 −6
Merged

Fix MLX multi-batch validation memory growth #32

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
321e82c
Fix MLX validation loss accumulation
yhn112 Mar 18, 2026
e17ed01
Log MLX validation progress
yhn112 Mar 18, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 14 additions & 6 deletions train_gpt_mlx.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -759,6 +759,7 @@ def eval_val(
base_bytes_lut: np.ndarray,
has_leading_space_lut: np.ndarray,
is_boundary_token_lut: np.ndarray,
log_fn: Callable[[str], None] | None = None,
) -> tuple[float, float]:
# Validation computes two metrics:
# - val_loss: token cross-entropy (natural log)
Expand All @@ -772,10 +773,11 @@ def eval_val(
)
val_batch_seqs = val_batch_tokens // args.train_seq_len
total_seqs = (val_tokens.size - 1) // args.train_seq_len
total_loss = mx.array(0.0, dtype=mx.float32)
total_batches = max((total_seqs + val_batch_seqs - 1) // val_batch_seqs, 1)
total_loss_sum = 0.0
total_tokens = 0.0
total_bytes = 0.0
for batch_seq_start in range(0, total_seqs, val_batch_seqs):
for batch_idx, batch_seq_start in enumerate(range(0, total_seqs, val_batch_seqs), start=1):
batch_seq_end = min(batch_seq_start + val_batch_seqs, total_seqs)
raw_start = batch_seq_start * args.train_seq_len
raw_end = batch_seq_end * args.train_seq_len + 1
Expand All @@ -785,7 +787,9 @@ def eval_val(
x = mx.array(x_np, dtype=mx.int32)
y = mx.array(y_np, dtype=mx.int32)
chunk_token_count = float(y.size)
total_loss = total_loss + compiled_loss(x, y).astype(mx.float32) * chunk_token_count
batch_loss = compiled_loss(x, y).astype(mx.float32)
mx.eval(batch_loss)
total_loss_sum += float(batch_loss.item()) * chunk_token_count
prev_ids = x_np.reshape(-1)
tgt_ids = y_np.reshape(-1)
bytes_np = base_bytes_lut[tgt_ids].astype(np.int16, copy=True)
Expand All @@ -794,9 +798,11 @@ def eval_val(
).astype(np.int16, copy=False)
total_tokens += chunk_token_count
total_bytes += float(bytes_np.astype(np.float64).sum())
total_loss = total_loss / total_tokens
mx.eval(total_loss)
val_loss = float(total_loss.item())
if log_fn is not None and total_batches > 1 and (
batch_idx == 1 or batch_idx == total_batches or batch_idx % 25 == 0
):
log_fn(f"val_progress:{batch_idx}/{total_batches}")
val_loss = total_loss_sum / total_tokens
bits_per_token = val_loss / math.log(2.0)
val_bpb = bits_per_token * (total_tokens / total_bytes)
return val_loss, val_bpb
Expand Down Expand Up @@ -999,6 +1005,7 @@ def log(msg: str, console: bool = True) -> None:
base_bytes_lut,
has_leading_space_lut,
is_boundary_token_lut,
log_fn=log,
)
train_time_ms += 1000.0 * (time.perf_counter() - t0)
if step % 25 == 0 or last_step:
Expand Down Expand Up @@ -1078,6 +1085,7 @@ def log(msg: str, console: bool = True) -> None:
base_bytes_lut,
has_leading_space_lut,
is_boundary_token_lut,
log_fn=log,
)
q_eval_ms = 1000.0 * (time.perf_counter() - q_t0)
log(f"final_int8_zlib_roundtrip val_loss:{q_val_loss:.4f} val_bpb:{q_val_bpb:.4f} eval_time:{q_eval_ms:.0f}ms")
Expand Down