J

jth01's activity

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

New activity in turboderp/Qwen3.8-Flash-Next-exl3 5 days ago

4bpw work on my 3090 + 64 RAM - excellence.

👍 10

21

#2 opened 6 days ago by

s1arsky

New activity in unsloth/granite-4.0-h-small-bnb-4bit 11 months ago

This isn't a 4 bit model

#1 opened 11 months ago by

jth01

New activity in unsloth/Qwen3-32B over 1 year ago

Base Model?

9

#2 opened over 1 year ago by

Downtown-Case

New activity in unsloth/Llama-4-Scout-17B-16E-Instruct-unsloth-bnb-4bit over 1 year ago

What's up with the minimal decrease in file size from fp16?

👀 1

1

#1 opened over 1 year ago by

jth01

New activity in unsloth/Llama-4-Scout-17B-16E-Instruct over 1 year ago

DOA

👍 1

15

#1 opened over 1 year ago by

MrDevolver

New activity in unsloth/Llama-4-Scout-17B-16E-Instruct-unsloth-bnb-4bit over 1 year ago

This looks like the base and not bnb-4bit?

4

#2 opened over 1 year ago by

bjj

New activity in unsloth/QwQ-32B over 1 year ago

I could be wrong, but I think the <think> tag needs to be removed from the last bit of the jinja template in the tokenizer_config.json

👍 1

3

#1 opened over 1 year ago by

jth01

New activity in qihoo360/TinyR1-32B-Preview over 1 year ago

Repeated Thinking Tags in Output Generation

10

#2 opened over 1 year ago by

xldistance

New activity in arcee-ai/Virtuoso-Medium-v2 over 1 year ago

Where are the benchmark scores?

2

#4 opened over 1 year ago by

rombodawg

New activity in mistralai/Mistral-Small-24B-Instruct-2501 over 1 year ago

This Mistral Small has FAR less knowledge than the last.

🔥 6

20

#5 opened over 1 year ago by

phil111

This Mistral Small has FAR less knowledge than the last.

🔥 6

20

#5 opened over 1 year ago by

phil111