mirror of
https://github.com/deepseek-ai/DeepInfer.git
synced 2026-10-28 23:07:09 +08:00
- nki_attention._chunk_attention: a prompt chunk run as a decode step gathers a dense compressed-KV scope once (every token's list is a prefix of the longest) and masks it per row; top-k and SWA scopes stay per row. The per-row gather was [128, 8192, 512] for a ratio-2 layer of the full model: 3.8 GB moved per call and a 3 GB NEFF (model code) that ran HBM out. - decode_part_timer: the decode timers are no-ops inside region graphs. Their f-string labels read layer_idx, so Dynamo guarded on it and retraced every layer instead of every layer kind. Tiny V4.1 on a NeuronCore, bucket 128: short 9/9, long 9/9 (logit cos >= 0.997). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> |
||
|---|---|---|
| .. | ||
| _C | ||
| distributed | ||
| engine | ||
| entrypoints | ||
| hugging_face | ||
| model_executor | ||
| neuron_ops | ||
| npu_ops | ||
| structure_output | ||
| ve | ||
| worker | ||
| __init__.py | ||
| attrdict.py | ||
| config.py | ||
| environ.py | ||
| load_class.py | ||
| logger.py | ||
| model_config_loader.py | ||
| multimodal_tokens.py | ||
| platform.py | ||
| py.typed | ||
| sampling_params.py | ||
| sequence.py | ||
| shared_memory.py | ||
| testing.py | ||
| timer.py | ||
| utils.py | ||