Zheng Han

traphix's activity

• All
• Models
• Datasets
• Spaces
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

New activity in Qwen/QwQ-32B 29 days ago

This model beats Qwen Max!

6

#33 opened about 1 month ago by

MrDevolver

New activity in Qwen/Qwen2.5-14B-Instruct-1M about 1 month ago

Does vllm 0.7.3 support this model？

#10 opened about 1 month ago by

traphix

New activity in cognitivecomputations/DeepSeek-V3-AWQ about 1 month ago

Is there any accuracy results comparing to original DeepSeek-V3？

#6 opened about 1 month ago by

traphix

New activity in cognitivecomputations/DeepSeek-R1-AWQ about 1 month ago

why "MLA is not supported with awq_marlin quantization. Disabling MLA." with 4090 * 32 (4 node / vllm 0.7.2)

3

#14 opened about 1 month ago by

FightLLM

Is there any accuracy results comparing to original DeepSeek-R1？

2

#15 opened about 1 month ago by

traphix

New activity in cognitivecomputations/DeepSeek-R1-AWQ about 2 months ago

Has anyone evaluated the performance of the AWQ version of the model on benchmarks?

4

#8 opened about 2 months ago by

liuqianchao

skips the thinking process

11

#5 opened about 2 months ago by

muzizon

Deployment framework

27

#2 opened 2 months ago by

xro7

New activity in cognitivecomputations/DeepSeek-V3-AWQ about 2 months ago

vllm support a100

17

#2 opened 3 months ago by

HuggingLianWang

New activity in neuralmagic/Qwen2.5-72B-quantized.w8a8 about 2 months ago

Any plans to quantize Qwen/Qwen2.5-72B-Instruct to w8a8？

#1 opened about 2 months ago by

traphix

New activity in neuralmagic/DeepSeek-Coder-V2-Instruct-FP8 7 months ago

Can it run on A100/A800 with VLLM?

3

#1 opened 8 months ago by

Parkerlambert123

liked a model 8 months ago

neuralmagic/DeepSeek-Coder-V2-Instruct-FP8

Text Generation • Updated Jul 22, 2024 • 203 • 7

upvoted a collection 8 months ago

INT8 LLMs for vLLM

Collection

Accurate INT8 quantized models by Neural Magic, ready for use with vLLM! • 50 items • Updated Sep 26, 2024 • 15

New activity in neuralmagic/DeepSeek-Coder-V2-Instruct-FP8 8 months ago

Quantize DeepSeek-Coder-V2-Instruct to W8A8(INT8)?

#2 opened 8 months ago by

traphix