Running
1
📈
AI & ML interests
None defined yet.
Recent Activity
View all activity
Qwen's activity
JustinLin610
updated
a
Space
about 1 hour ago
Post
426
LLaVA-Mini🔥 A efficient multimodal model for image and video understanding released by Chinese Academy of Sciences
Model: ICTNLP/llava-mini-llama-3.1-8b
Paper: LLaVA-Mini: Efficient Image and Video Large Multimodal Models with One Vision Token (2501.03895)
✨ Matches LLaVA-v1.5 using just 1 vision token
✨ Delivers <40ms response time
✨ Reduces vision tokens while maintaining strong visual understanding
Model: ICTNLP/llava-mini-llama-3.1-8b
Paper: LLaVA-Mini: Efficient Image and Video Large Multimodal Models with One Vision Token (2501.03895)
✨ Matches LLaVA-v1.5 using just 1 vision token
✨ Delivers <40ms response time
✨ Reduces vision tokens while maintaining strong visual understanding
bluelike
authored
a
paper
1 day ago
JustinLin610
authored
a
paper
1 day ago
qwen-turbo-1101
7
#1 opened about 2 months ago
by
LaferriereJC
alielfilali01
posted
an
update
4 days ago
Post
1703
3C3H AraGen Leaderboard welcomes today
deepseek-ai/DeepSeek-V3 and 12 other models (including the late gpt-3.5 💀) to the ranking of best LLMs in Arabic !
Observations:
- DeepSeek-v3 ranked 3rd and only Open model among the top 5 !
- A 14B open model ( Qwen/Qwen2.5-14B-Instruct) outperforms gpt-3.5-turbo-0125 (from last year). This shows how much we came in advancing and supporting Arabic presence within the LLM ecosystem !
- Contrary to what observed in likelihood-acc leaderboards (like OALL/Open-Arabic-LLM-Leaderboard) further finetuned models like maldv/Qwentile2.5-32B-Instruct actually decreased the performance compared to the original model Qwen/Qwen2.5-32B-Instruct.
It's worth to note that the decrease is statiscally insignificant which imply that at best, the out-domain finetuning do not really hurts the model original capabilities acquired during pretraining.
Previous work addressed this (finetuning VS pretraining) but more investigation in this regard is required (any PhDs here ? This could be your question ...)
Check out the latest rankings: inceptionai/AraGen-Leaderboard
Observations:
- DeepSeek-v3 ranked 3rd and only Open model among the top 5 !
- A 14B open model ( Qwen/Qwen2.5-14B-Instruct) outperforms gpt-3.5-turbo-0125 (from last year). This shows how much we came in advancing and supporting Arabic presence within the LLM ecosystem !
- Contrary to what observed in likelihood-acc leaderboards (like OALL/Open-Arabic-LLM-Leaderboard) further finetuned models like maldv/Qwentile2.5-32B-Instruct actually decreased the performance compared to the original model Qwen/Qwen2.5-32B-Instruct.
It's worth to note that the decrease is statiscally insignificant which imply that at best, the out-domain finetuning do not really hurts the model original capabilities acquired during pretraining.
Previous work addressed this (finetuning VS pretraining) but more investigation in this regard is required (any PhDs here ? This could be your question ...)
Check out the latest rankings: inceptionai/AraGen-Leaderboard
Post
4313
Switching to
I posted a poll on twitter, and others have mentioned the interest in me using the convention of including the author name in the model path when I upload.
It has a couple advantages, first and foremost of course is ensuring clarity of who uploaded the original model (did Qwen upload Qwen2.6? Or did someone fine tune Qwen2.5 and named it 2.6 for fun?)
The second thing is that it avoids collisions, so if multiple people upload the same model and I try to quant them both, I would normally end up colliding and being unable to upload both
I'll be implementing the change next week, there are just two final details I'm unsure about:
First, should the files also inherit the author's name?
Second, what to do in the case that the author name + model name pushes us past the character limit?
Haven't yet decided how to handle either case, so feedback is welcome, but also just providing this as a "heads up"
author_model-nameI posted a poll on twitter, and others have mentioned the interest in me using the convention of including the author name in the model path when I upload.
It has a couple advantages, first and foremost of course is ensuring clarity of who uploaded the original model (did Qwen upload Qwen2.6? Or did someone fine tune Qwen2.5 and named it 2.6 for fun?)
The second thing is that it avoids collisions, so if multiple people upload the same model and I try to quant them both, I would normally end up colliding and being unable to upload both
I'll be implementing the change next week, there are just two final details I'm unsure about:
First, should the files also inherit the author's name?
Second, what to do in the case that the author name + model name pushes us past the character limit?
Haven't yet decided how to handle either case, so feedback is welcome, but also just providing this as a "heads up"
- 3 replies
Post
2237
Excited to see Alibaba DAMO Academy release a multimodel dataset for vision language pretraining on the hub🔥
Paper: 2.5 Years in Class: A Multimodal Textbook for Vision-Language Pretraining (2501.00958)
Dataset: DAMO-NLP-SG/multimodal_textbook
✨ 6.5M images + 0.8B text from 22k hours of instructional videos
✨ Covers subjects like math, physics, and chemistry
✨ Apache 2.0
Paper: 2.5 Years in Class: A Multimodal Textbook for Vision-Language Pretraining (2501.00958)
Dataset: DAMO-NLP-SG/multimodal_textbook
✨ 6.5M images + 0.8B text from 22k hours of instructional videos
✨ Covers subjects like math, physics, and chemistry
✨ Apache 2.0
jx-yang
authored
a
paper
5 days ago
bzheng
authored
a
paper
5 days ago
Losin94
authored
a
paper
5 days ago
JustinLin610
authored
a
paper
5 days ago
Update README.md
#5 opened 6 days ago
by
quanshr
Create test.json
#4 opened 6 days ago
by
quanshr
xingzhang
authored
5
papers
8 days ago
Qwen Technical Report
Paper
•
2309.16609
•
Published
•
35
PolyLM: An Open Source Polyglot Large Language Model
Paper
•
2307.06018
•
Published
•
25
Qwen2 Technical Report
Paper
•
2407.10671
•
Published
•
160
Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement
Paper
•
2409.12122
•
Published
•
3
Qwen2.5 Technical Report
Paper
•
2412.15115
•
Published
•
339