squeeze-ai-lab commited on
Commit
a572e1b
1 Parent(s): d2ee9b7

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -5
README.md CHANGED
@@ -17,7 +17,7 @@ language:
17
  <p align="center">
18
  <a href="https://github.com/SqueezeAILab/TinyAgent/raw/main/TinyAgent.zip">Get the desktop app</a>‎ ‎
19
  |
20
- <a href="https://bair.berkeley.edu/blog/2024/05/28/tiny-agent">Read the blog post</a>
21
  </p>
22
 
23
  ![Thumbnail](https://cdn-uploads.huggingface.co/production/uploads/648903e1ce7b9a2abe3511aa/a1YuQosFiJQJ_7Ejribrd.png)
@@ -44,7 +44,7 @@ Please see our [Github](https://github.com/SqueezeAILab/TinyAgent) for details o
44
  ## Training Details
45
 
46
  **Dataset:**
47
- We curated a [dataset](https://huggingface.co/datasets/squeeze-ai-lab/TinyAgent-dataset) of **40,000** real-life use cases. We use GPT-3.5-Turbo to generate real-world instructions. These are then used to obtain synthetic execution plans using GPT-4-Turbo. Please check out our blog post for more details on our dataset.
48
 
49
  **Fine-tuning Procedure:**
50
  TinyAgent models are fine-tuned from base models. Below is a table of each TinyAgent model with its base counterpart
@@ -57,13 +57,13 @@ TinyAgent models are fine-tuned from base models. Below is a table of each TinyA
57
  | TinyAgent-1.1B + ToolRAG / [[hf](https://huggingface.co/squeeze-ai-lab/TinyAgent-1.1B)] [[gguf](https://huggingface.co/squeeze-ai-lab/TinyAgent-1.1B-GGUF)] | **80.06%** |
58
  | TinyAgent-7B + ToolRAG / [[hf](https://huggingface.co/squeeze-ai-lab/TinyAgent-7B)] [[gguf](https://huggingface.co/squeeze-ai-lab/TinyAgent-7B-GGUF)] | **84.95%** |
59
 
60
- Using the synthetic data generation process described above, we use parameter-efficient fine-tuning with LoRA to fine-tune the base models for 3 epochs. Please check out our blog post for more details on our fine-tuning procedure.
61
 
62
  ### 🛠️ ToolRAG
63
 
64
- When faced with challenging tasks, SLM agents require appropriate tools and in-context examples to guide them. If the model sees irrelevant examples, it can hallucinate. Likewise, if the model sees the descriptions of the tools that it doesn’t need, it usually gets confused, and these tools take up unnecessary prompt space. To tackle this, TinyAgent uses ToolRAG to retrieve the best tools and examples suited for a given query. This process has minimal latency and increases the accuracy of TinyAgent substantially. Please take a look at our blog post and our [ToolRAG model](https://huggingface.co/squeeze-ai-lab/TinyAgent-ToolRAG) for more details.
65
 
66
  ## Links
67
- **Blog Post**:
68
 
69
  **Github:** https://github.com/SqueezeAILab/TinyAgent
 
17
  <p align="center">
18
  <a href="https://github.com/SqueezeAILab/TinyAgent/raw/main/TinyAgent.zip">Get the desktop app</a>‎ ‎
19
  |
20
+ <a href="https://bair.berkeley.edu/blog/2024/05/29/tiny-agent/">Read the blog post</a>
21
  </p>
22
 
23
  ![Thumbnail](https://cdn-uploads.huggingface.co/production/uploads/648903e1ce7b9a2abe3511aa/a1YuQosFiJQJ_7Ejribrd.png)
 
44
  ## Training Details
45
 
46
  **Dataset:**
47
+ We curated a [dataset](https://huggingface.co/datasets/squeeze-ai-lab/TinyAgent-dataset) of **40,000** real-life use cases. We use GPT-3.5-Turbo to generate real-world instructions. These are then used to obtain synthetic execution plans using GPT-4-Turbo. Please check out our [blog post](https://bair.berkeley.edu/blog/2024/05/29/tiny-agent/) for more details on our dataset.
48
 
49
  **Fine-tuning Procedure:**
50
  TinyAgent models are fine-tuned from base models. Below is a table of each TinyAgent model with its base counterpart
 
57
  | TinyAgent-1.1B + ToolRAG / [[hf](https://huggingface.co/squeeze-ai-lab/TinyAgent-1.1B)] [[gguf](https://huggingface.co/squeeze-ai-lab/TinyAgent-1.1B-GGUF)] | **80.06%** |
58
  | TinyAgent-7B + ToolRAG / [[hf](https://huggingface.co/squeeze-ai-lab/TinyAgent-7B)] [[gguf](https://huggingface.co/squeeze-ai-lab/TinyAgent-7B-GGUF)] | **84.95%** |
59
 
60
+ Using the synthetic data generation process described above, we use parameter-efficient fine-tuning with LoRA to fine-tune the base models for 3 epochs. Please check out our [blog post](https://bair.berkeley.edu/blog/2024/05/29/tiny-agent/) for more details on our fine-tuning procedure.
61
 
62
  ### 🛠️ ToolRAG
63
 
64
+ When faced with challenging tasks, SLM agents require appropriate tools and in-context examples to guide them. If the model sees irrelevant examples, it can hallucinate. Likewise, if the model sees the descriptions of the tools that it doesn’t need, it usually gets confused, and these tools take up unnecessary prompt space. To tackle this, TinyAgent uses ToolRAG to retrieve the best tools and examples suited for a given query. This process has minimal latency and increases the accuracy of TinyAgent substantially. Please take a look at our [blog post](https://bair.berkeley.edu/blog/2024/05/29/tiny-agent/) and our [ToolRAG model](https://huggingface.co/squeeze-ai-lab/TinyAgent-ToolRAG) for more details.
65
 
66
  ## Links
67
+ **Blog Post**: https://bair.berkeley.edu/blog/2024/05/29/tiny-agent/
68
 
69
  **Github:** https://github.com/SqueezeAILab/TinyAgent