Skip to content

NVFP4 quantization: streaming loader, fused MoE experts (Qwen + Llama… #45537

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

Closed
ddreeselogs wants to merge 1 commit into from
Closed

NVFP4 quantization: streaming loader, fused MoE experts (Qwen + Llama… #45537

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
1 change: 1 addition & 0 deletions .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -184,3 +184,4 @@ test-results/
# AI agent local setup artifacts
/.agents/skills
/.claude/skills
*.bak
7 changes: 7 additions & 0 deletions src/transformers/core_model_loading.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1211,7 +1211,14 @@ def convert_and_load_state_dict_in_model(
pattern_to_converter = {k: converter for converter in converters for k in converter.source_patterns}

state_dict = sorted(state_dict.items(), key=lambda kv: dot_natural_key(kv[0]))
import gc as _gc
_tensor_count = 0
for original_key, tensor in state_dict:
_tensor_count += 1
if _tensor_count % 1000 == 0:
_gc.collect()
if torch.cuda.is_available():
torch.cuda.empty_cache()
# 1. Rename the key according to all renaming pattern and optional weight converter patterns
renamed_key, source_pattern = rename_source_key(
original_key, renamings, converters, prefix, meta_model_state_dict
Expand Down
Loading
Loading