Jon Doe

bullpoint's activity

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

New activity in local-inference-lab/GLM-5.3-NVFP4 about 2 months ago

Exclude BF16 MTP layer 78 from ModelOpt FP4

#2 opened about 2 months ago by

bullpoint

Upload chat_template.jinja

#1 opened about 2 months ago by

bullpoint

New activity in Qwen/Qwen3.5-397B-A17B 6 months ago

Qwen3.6 397b

3

#75 opened 7 months ago by

zenmagnets

New activity in moonshotai/Kimi-K2.6 6 months ago

Fast tokenizer token-ID parity with previous slow tokenizer?

3

#39 opened 6 months ago by

bullpoint

New activity in bullpoint/Qwen3-Coder-Next-AWQ-4bit 8 months ago

vllm says it's not AWQ

4

#2 opened 9 months ago by

jcowles

Works great! Thannk you!

3

#1 opened 9 months ago by

Max-NexaEthos

New activity in nvidia/Qwen3.5-397B-A17B-NVFP4 8 months ago

Getting nvidia/Qwen3.5-397B-A17B-NVFP4 running with SGLang (requires transformers v5) on RTX PRO 6000 (blackwell) CUDA 12.9

🔥 1

6

#1 opened 9 months ago by

bullpoint

New activity in vincentzed-hf/Qwen3.5-397B-A17B-NVFP4 9 months ago

Anyone try this on 4x RTX 6000 Pro yet?

52

#1 opened 9 months ago by

zenmagnets

New activity in bullpoint/GLM-4.6-AWQ 11 months ago

endless response

7

#4 opened 11 months ago by

ramidahbash

GLM-4.6-FP8 - 55 tokens/sec on 4x RTX 6000 PRO

6

#2 opened about 1 year ago by

festr2

Speed measured on 2 x H200 / 4 x RTX 6000 Pro / 4 x A100

1

#3 opened 11 months ago by

HristoTodorov

New activity in bullpoint/GLM-4.6-AWQ about 1 year ago

llm-compressor version

1

#1 opened about 1 year ago by

bullerwins

New activity in microsoft/phi-4 almost 2 years ago

Suggested tokenizer changes by Unsloth.ai

🔥 13

7

#21 opened almost 2 years ago by

gugarosa