support W4afp8 quant in v3.1 #337
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: bruce.xu <[email protected]>
Bruce-x-1997
force-pushed
the
bruce-v3.1-w4afp8
branch
from
b3f4b5e to
8824e8b
Compare
|
@cjluo-nv please help review it, thanks |
cjluo-nv
reviewed
Sep 25, 2025
| assert weight_quantizer is None | ||
| assert act_quantizer is None | ||
| x, scale = act_quant(x, block_size) | ||
| x, scale = act_quant(x, block_size, scale_fmt) |
There was a problem hiding this comment.
could you walk through how updating the scale will give you W4A8?
In this case, what's the 4bit weight? Is it NVFP4 or INT4?
There was a problem hiding this comment.
sorry, I don't notice the comment before.
in my case, 4bit weight means int4.
in 3.1 case, when I want to quant the model, there is a interface mismatch with deepseek-v3.git
so I fix it @cjluo-nv
cjluo-nv
reviewed
Sep 25, 2025
| nvcr.io/nvidia/tensorrt-llm/release | ||
| ``` | ||
| then we can operate modelopt in the docker pod as (Trtllm example)[https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/models/core/deepseek_v3/README.md?plain=1] | ||
| but we should notice that just using the latest DeepSeek-V3.git is ok, because there is a dtype bug in bias proto at commit 1398800. |
There was a problem hiding this comment.
would you recommend using a more recent commit?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
support w4afp8 quant in v3.1(ue8m0)
Usage
just set gemm_impl to fp8 and use config_v3.1.json
Testing
after applying this patch, our model(3.1 using w4afp8) could reach aime25 50% and aime24 60%
Before your PR is "Ready for review"
Additional Information