anakin87 (Stefano Fiorucci)

Posts 8

Post

1473

🕵🏻 𝐀𝐠𝐞𝐧𝐭𝐢𝐜 𝐑𝐀𝐆 𝐰𝐢𝐭𝐡 🦙 𝐋𝐥𝐚𝐦𝐚 3.2

I was excited to explore Llama 3.2, but as a simple 🇪🇺 EU guy, I don't have access to Meta's multimodal models 😿

🤔 So I thought: why not challenge the small 3B text model with Agentic RAG?

🎯 The plan:
- Build a system that tries to answer questions using a knowledge base.
- If the documents don't contain the answer, use Web search for additional context.

Check out my experimental notebook here: 📓 https://colab.research.google.com/github/deepset-ai/haystack-cookbook/blob/main/notebooks/llama32_agentic_rag.ipynb

My stack:
🏗️ haystack (https://haystack.deepset.ai/): open-source LLM orchestration framework
🦙 meta-llama/Llama-3.2-3B-Instruct
🦆🌐 free DuckDuckGo API, integrated with Haystack

✨ 𝘛𝘩𝘦 𝘳𝘦𝘴𝘶𝘭𝘵𝘴? 𝘌𝘯𝘤𝘰𝘶𝘳𝘢𝘨𝘪𝘯𝘨 - 𝘢 𝘧𝘦𝘸 𝘮𝘰𝘯𝘵𝘩𝘴 𝘢𝘨𝘰, 𝘵𝘩𝘪𝘴 𝘭𝘦𝘷𝘦𝘭 𝘰𝘧 𝘱𝘦𝘳𝘧𝘰𝘳𝘮𝘢𝘯𝘤𝘦 𝘧𝘳𝘰𝘮 𝘢 𝘴𝘮𝘢𝘭𝘭 𝘮𝘰𝘥𝘦𝘭 𝘸𝘰𝘶𝘭𝘥'𝘷𝘦 𝘣𝘦𝘦𝘯 𝘶𝘯𝘵𝘩𝘪𝘯𝘬𝘢𝘣𝘭𝘦!
This probably reflects the impressive IFEval score of the model (comparable to Llama 3.1 8B).

Post

1039

𝐌𝐲 𝐟𝐢𝐫𝐬𝐭 𝐜𝐨𝐦𝐦𝐮𝐧𝐢𝐭𝐲 𝐚𝐫𝐭𝐢𝐜𝐥𝐞! 𝐒𝐞𝐥𝐞𝐜𝐭𝐢𝐯𝐞 𝐟𝐢𝐧𝐞-𝐭𝐮𝐧𝐢𝐧𝐠 𝐰𝐢𝐭𝐡 𝐒𝐩𝐞𝐜𝐭𝐫𝐮𝐦 🎯

Full walkthrough on how to get started with Spectrum and TRL for efficient fine-tuning.
📔 👣 https://huggingface.co/blog/anakin87/spectrum

---

Looking to fine-tune Language Models efficiently and save on computational resources?

One popular method is QLoRa, which quantizes the original model and trains low-rank adapters on top.
It's quite effective and uses less GPU than full fine-tuning.

However, QLoRa applies Low-Rank Adaptation uniformly across the entire model.

What if we could identify the most informative layers and only fine-tune those? 🤔

This is exactly what Spectrum does! 👇

🔬 Spectrum analyzes the weight matrices for all layers in a Language Model and calculates a Signal to Noise Ratio (SNR) for each one.
(It uses Random Matrix Theory and Marchenko-Pastur distribution to distinguish signal from noise.)

🎯 Based on a chosen percentage (say, 25%), Spectrum selects the most informative layers of each type (mlp.down_proj, self_attn.o_proj, etc.).

You can then ❄️ freeze the rest of the model and focus your 🏋️‍♂️ training on the chosen layers.

🏆 Results/Evaluation
- Spectrum is competitive with full fine-tuning and beats QLoRA on benchmarks.
- While QLoRA is more memory-efficient on a single GPU, Spectrum shines in distributed training setups.
- Great models trained with Spectrum: Dolphin models, Llama 3.1 Storm, numerous models by VAGO Solutions...

---

For a practical guide, check out the article above.

View all posts