d-flex flex-column width-full">

[Fix] Limit static FP4 linear kv_b_proj post-processing

ffe41ae

Select commit

Loading

Failed to load commit list.

Uh oh!

There was an error while loading. Please reload this page.

Sign in for the full log view

Pre Checkin on: pull_request

Check Code Style with Black

Check Code Style with Ruff

ATOM Test on: pull_request 26

Check Pre Checkin Signal

Load test model configs

Download aiter wheel

Accuracy (Meta-Llama-3-8B-Instruct, meta-llama/Meta-Llama-3-8B-Instruct, --kv_cache_dtype fp8 --g...

Accuracy (DeepSeek-R1-0528, deepseek-ai/DeepSeek-R1-0528, --kv_cache_dtype fp8 -tp 8, atom-mi355-...

Accuracy (DeepSeek-V4-Pro, deepseek-ai/DeepSeek-V4-Pro, --kv_cache_dtype fp8 -tp 8, AITER_BF16_FP...

Accuracy (DeepSeek-V4-Pro MTP, deepseek-ai/DeepSeek-V4-Pro, --kv_cache_dtype fp8 -tp 8 --method m...

Accuracy (gpt-oss-120b, openai/gpt-oss-120b, --kv_cache_dtype fp8 --gpu-memory-utilization 0.5, l...

Accuracy (DeepSeek-R1-0528-FP4 MTP, amd/DeepSeek-R1-0528-MXFP4-MTP-MoEFP4, --kv_cache_dtype fp8 -...

Accuracy (Qwen3-Next-80B-A3B-Thinking, Qwen/Qwen3-Next-80B-A3B-Thinking, -tp 8 --kv_cache_dtype f...

Accuracy (Kimi-K2.5-MXFP4, amd/Kimi-K2.5-MXFP4, --kv_cache_dtype fp8 -tp 4 --trust-remote-code, H...

Accuracy (GLM-5.1-FP8, zai-org/GLM-5.1-FP8, --kv_cache_dtype fp8 -tp 8 --default-chat-template-kw...

Accuracy (GLM-5.1-MXFP4 MTP, amd/GLM-5.1-MXFP4, --kv_cache_dtype fp8 -tp 4 --default-chat-templat...

Accuracy (GLM-5.1-MXFP4, amd/GLM-5.1-MXFP4, --kv_cache_dtype fp8 -tp 4 --default-chat-template-kw...

Accuracy (Qwen3.5-397B-A17B-MXFP4, amd/Qwen3.5-397B-A17B-MXFP4, --kv_cache_dtype fp8 -tp 4, ATOM_...

Update accuracy dashboard

ATOM SGLang Test on: pull_request 4

Check Pre Checkin Signal

Download aiter wheel

Accuracy (DeepSeek-R1-FP8 TP4)

Accuracy (DeepSeek-R1-FP4 TP4)

Accuracy (Qwen3.5-35B-A3B-FP8 TP2)

Build and Deploy Documentation on: pull_request 1

build-docs

deploy-docs

ATOM vLLM Test on: pull_request 9

Check Pre Checkin Signal

Download aiter wheel

Accuracy (DeepSeek-R1-FP8 TP8)

Accuracy (gpt-oss-120b TP1)

Accuracy (Kimi-K2-Thinking-MXFP4 TP4)

Accuracy (Qwen3.5-35B-A3B-FP8 TP2)

ATOM vLLM Test on: pull_request

Check Pre Checkin Signal

Download aiter wheel

Accuracy (${{ matrix.display_name }})

ATOM SGLang Test on: pull_request

Check Pre Checkin Signal

Download aiter wheel

Accuracy (${{ matrix.model_name }})

The logs for this run have expired and are no longer available.

Loading

[Fix] Limit static FP4 linear kv_b_proj post-processing

Uh oh!

The logs for this run have expired and are no longer available.

Re-running jobs...