None

Thireus's activity

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

liked a model 3 days ago

DavidAU/Qwen3-4B-Q8_0-64k-128k-256k-context-GGUF

Text Generation • Updated 12 days ago • 280 • 1

New activity in unsloth/Qwen3-32B-GGUF 4 days ago

Potentially still broken?

3

13

#8 opened 12 days ago by

qenthousiast

liked a model 4 days ago

unsloth/Qwen3-32B-128K-GGUF

Text Generation • Updated 4 days ago • 17.5k • 18

New activity in kalomaze/Qwen3-16B-A3B 5 days ago

Brainstorming

4

5

#6 opened 11 days ago by

Downtown-Case

New activity in unsloth/Qwen3-32B-128K-GGUF 5 days ago

UD version not doing great with YaRN compared to non-UD of the same size

4

#4 opened 13 days ago by

Thireus

New activity in unsloth/Qwen3-32B-GGUF 5 days ago

PPL vs model size - safe to assume larger size == better accuracy regardless of UD vs non-UD?

3

#6 opened 15 days ago by

Thireus

New activity in Qwen/Qwen3-32B 9 days ago

Potential issue with large context sizes - can someone confirm?

13

#18 opened 14 days ago by

Thireus

New activity in Qwen/Qwen3-235B-A22B 10 days ago

Qwen is loosing broad knowledge since Qwen2.

10

14

#16 opened 15 days ago by

phil111

New activity in unsloth/Qwen3-235B-A22B-128K-GGUF 10 days ago

YaRN not enabled correctly

1

1

#3 opened 15 days ago by

CISCai

New activity in unsloth/Qwen3-0.6B-GGUF 10 days ago

Any chance of a 128k version so we can use it as a draft model for the larger 128k models?

3

8

#3 opened 14 days ago by

smcleod

New activity in unsloth/Qwen3-16B-A3B-GGUF 11 days ago

Could we have the 128K variants as well please?

1

#1 opened 11 days ago by

Thireus

New activity in kalomaze/Qwen3-16B-A3B 11 days ago

Context size? YaRN still supported?

2

#3 opened 11 days ago by

Thireus

liked a model 11 days ago

unsloth/Qwen3-16B-A3B-GGUF

Updated 11 days ago • 3.55k • 7

liked 2 models 15 days ago

Qwen/Qwen3-32B

Text Generation • Updated 15 days ago • 347k • • 300

Qwen/Qwen3-30B-A3B

Text Generation • Updated 14 days ago • 178k • • 537

reacted to eaddario's post with 👍 16 days ago

view post

Post

2249

Until recently, watt-ai/watt-tool-70B was the best performing model in the Berkeley Function-Calling Leaderboard (https://gorilla.cs.berkeley.edu/leaderboard.html), which evaluates LLM's ability to call functions (tools) accurately. The top spot now belongs to Salesforce/Llama-xLAM-2-70b-fc-r and by a quite wide margin!

Layer-wise quantized versions for both models are available at eaddario/Llama-xLAM-2-8b-fc-r-GGUF and eaddario/Watt-Tool-8B-GGUF

See translation

🚀

3

3

👍

2

2

+

Reply

liked 4 models about 1 month ago

ordis-co-ltd/Qwen2.5-VL-72B-Instruct_exl2_6.0bpw

Image-Text-to-Text • Updated Feb 10 • 21 • 2

city96/Wan2.1-I2V-14B-720P-gguf

Image-to-Video • Updated Feb 28 • 58.6k • 100

city96/Wan2.1-T2V-14B-gguf

Text-to-Video • Updated Feb 26 • 45.1k • 135

hexgrad/Kokoro-82M

Text-to-Speech • Updated Apr 10 • 1.87M • • 4.3k