agentica-org/DeepCoder-14B-Preview
Text Generation
•
Updated
•
37.1k
•
604
hassenhamdi
hassenhamdi
AI & ML interests
None yet
Recent Activity
liked
a model
15 days ago
agentica-org/DeepCoder-14B-Preview
liked
a model
16 days ago
IndexTeam/Index-TTS
Organizations
hassenhamdi's activity
liked
a
model
15 days ago
liked
a
model
16 days ago
reacted to
sr-rai's
post with 🤗
18 days ago
Post
2638
ExLlamaV3 is out. And it introduces EXL3 - a new SOTA quantization format!
"The conversion process is designed to be simple and efficient and requires only an input model (in HF format) and a target bitrate. By computing Hessians on the fly and thanks to a fused Viterbi kernel, the quantizer can convert a model in a single step, taking a couple of minutes for smaller models, up to a few hours for larger ones (70B+) (on a single RTX 4090 or equivalent GPU.)"
Repo: https://github.com/turboderp-org/exllamav3
"The conversion process is designed to be simple and efficient and requires only an input model (in HF format) and a target bitrate. By computing Hessians on the fly and thanks to a fused Viterbi kernel, the quantizer can convert a model in a single step, taking a couple of minutes for smaller models, up to a few hours for larger ones (70B+) (on a single RTX 4090 or equivalent GPU.)"
Repo: https://github.com/turboderp-org/exllamav3
- 1 reply
it will be great addition if it does maybe with an editor like reddit 🤖 post .
fix the link to github it has ')**' at the end.
Also it appears that huggingface posts don't handle markdowns.
liked
2
models
25 days ago
reacted to
MohamedRashad's
post with 🤝
25 days ago
Post
2445
I collected the recitations of the holy quran from 20 different reciters and uploaded the full dataset here:
MohamedRashad/Quran-Recitations
Check it out 🥷
MohamedRashad/Quran-Recitations
Check it out 🥷
liked
a
Space
about 1 month ago
reacted to
singhsidhukuldeep's
post with 🧠
about 2 months ago
Post
3510
O1 Embedder: Transforming Retrieval Models with Reasoning Capabilities
Researchers from University of Science and Technology of China and Beijing Academy of Artificial Intelligence have developed a novel retrieval model that mimics the slow-thinking capabilities of reasoning-focused LLMs like OpenAI's O1 and DeepSeek's R1.
Unlike traditional embedding models that directly match queries with documents, O1 Embedder first generates thoughtful reflections about the query before performing retrieval. This two-step process significantly improves performance on complex retrieval tasks, especially those requiring intensive reasoning or zero-shot generalization to new domains.
The technical implementation is fascinating:
- The model integrates two essential functions: Thinking and Embedding
- It uses an "Exploration-Refinement" data synthesis workflow where initial thoughts are generated by an LLM and refined by a retrieval committee
- A multi-task training method fine-tunes a pre-trained LLM to generate retrieval thoughts via behavior cloning while simultaneously learning embedding capabilities through contrastive learning
- Memory-efficient joint training enables both tasks to share encoding results, dramatically increasing batch size
The results are impressive - O1 Embedder outperforms existing methods across 12 datasets in both in-domain and out-of-domain scenarios. For example, it achieves a 3.9% improvement on Natural Questions and a 3.0% boost on HotPotQA compared to models without thinking capabilities.
This approach represents a significant paradigm shift in retrieval technology, bridging the gap between traditional dense retrieval and the reasoning capabilities of large language models.
What do you think about this approach? Could "thinking before retrieval" transform how we build search systems?
Researchers from University of Science and Technology of China and Beijing Academy of Artificial Intelligence have developed a novel retrieval model that mimics the slow-thinking capabilities of reasoning-focused LLMs like OpenAI's O1 and DeepSeek's R1.
Unlike traditional embedding models that directly match queries with documents, O1 Embedder first generates thoughtful reflections about the query before performing retrieval. This two-step process significantly improves performance on complex retrieval tasks, especially those requiring intensive reasoning or zero-shot generalization to new domains.
The technical implementation is fascinating:
- The model integrates two essential functions: Thinking and Embedding
- It uses an "Exploration-Refinement" data synthesis workflow where initial thoughts are generated by an LLM and refined by a retrieval committee
- A multi-task training method fine-tunes a pre-trained LLM to generate retrieval thoughts via behavior cloning while simultaneously learning embedding capabilities through contrastive learning
- Memory-efficient joint training enables both tasks to share encoding results, dramatically increasing batch size
The results are impressive - O1 Embedder outperforms existing methods across 12 datasets in both in-domain and out-of-domain scenarios. For example, it achieves a 3.9% improvement on Natural Questions and a 3.0% boost on HotPotQA compared to models without thinking capabilities.
This approach represents a significant paradigm shift in retrieval technology, bridging the gap between traditional dense retrieval and the reasoning capabilities of large language models.
What do you think about this approach? Could "thinking before retrieval" transform how we build search systems?
upvoted
a
collection
about 2 months ago
Would like to see performance on well known benchmark datasets.
reacted to
wassemgtk's
post with 🧠
about 2 months ago
Post
1879
# GESAL: Real-Time Adaptation for LLMs
We’re excited to unveil **Graph-Enhanced Singular Adaptive Learning (GESAL)**, a framework that lets LLMs like
🔗 **Code**: [https://github.com/writer/AI-Adaptive-Learning-GESAL](https://github.com/writer/AI-Adaptive-Learning-GESAL)
---
## Why GESAL?
Static LLMs struggle to adapt without heavy retraining. GESAL solves this with:
- **SVF**: Adapts weights via \( W' = U (\Sigma \cdot z) V^T \), using few parameters.
- **Graph Memory**: Stores adaptations in nodes for scalability.
- **RL**: Updates via \( J(z) = \mathbb{E}[\log \pi_z(y|x) r] \) based on feedback.
---
## How It Works
Ask "How many R’s in ‘strawberry’?" If it says "2" and you say "no," GESAL learns to say "3" next time, avoiding repeats.
---
## Try It
Built with Hugging Face’s
Needs a Hugging Face token for Llama-3.2-1B.
---
## Results
GESAL hits 95% accuracy after 5 feedbacks vs. LoRA’s 70%. It’s efficient (~0.5M params) and scalable.
We’re excited to unveil **Graph-Enhanced Singular Adaptive Learning (GESAL)**, a framework that lets LLMs like
meta-llama/Llama-3.2-1B adapt in real time using user feedback. Check out the code and white paper on GitHub!🔗 **Code**: [https://github.com/writer/AI-Adaptive-Learning-GESAL](https://github.com/writer/AI-Adaptive-Learning-GESAL)
---
## Why GESAL?
Static LLMs struggle to adapt without heavy retraining. GESAL solves this with:
- **SVF**: Adapts weights via \( W' = U (\Sigma \cdot z) V^T \), using few parameters.
- **Graph Memory**: Stores adaptations in nodes for scalability.
- **RL**: Updates via \( J(z) = \mathbb{E}[\log \pi_z(y|x) r] \) based on feedback.
---
## How It Works
Ask "How many R’s in ‘strawberry’?" If it says "2" and you say "no," GESAL learns to say "3" next time, avoiding repeats.
---
## Try It
Built with Hugging Face’s
transformers:pip install transformers torch numpy
python Adaptive_Learning_(GESAL).pyNeeds a Hugging Face token for Llama-3.2-1B.
---
## Results
GESAL hits 95% accuracy after 5 feedbacks vs. LoRA’s 70%. It’s efficient (~0.5M params) and scalable.
-
- 15 replies
reacted to
freddyaboulton's
post with 🚀
about 2 months ago
Post
3259
Getting WebRTC and Websockets right in python is very tricky. If you've tried to wrap an LLM in a real-time audio layer then you know what I'm talking about.
That's where FastRTC comes in! It makes WebRTC and Websocket streams super easy with minimal code and overhead.
Check out our org: hf.co/fastrtc
That's where FastRTC comes in! It makes WebRTC and Websocket streams super easy with minimal code and overhead.
Check out our org: hf.co/fastrtc
reacted to
KonradSzafer's
post with 👀
about 2 months ago
Post
1875
I’ve been experimenting with a “Tech Tree” to make ML research more systematic and transparent—turns out it helped me spot hidden interactions between experiments and share progress more easily. I wrote a short blog post with examples and insights!
KonradSzafer/tech_tree_blog
liked
a
Space
about 2 months ago
upvoted
a
paper
about 2 months ago
liked
3
models
about 2 months ago
