Upload README.md
Browse files- temp_model/README.md +40 -0
temp_model/README.md
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# BitAgent Tool-Calling Model
|
| 2 |
+
|
| 3 |
+
This model is specifically trained for tool calling tasks with special handling for distance calculations.
|
| 4 |
+
|
| 5 |
+
## Model Description
|
| 6 |
+
|
| 7 |
+
This model is designed to handle tool calling tasks with specific emphasis on:
|
| 8 |
+
- Parameter handling for distance calculations
|
| 9 |
+
- Correct argument ordering for origin/destination pairs
|
| 10 |
+
- Function call formatting
|
| 11 |
+
|
| 12 |
+
## Usage
|
| 13 |
+
|
| 14 |
+
```python
|
| 15 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
| 16 |
+
|
| 17 |
+
# Load model and tokenizer
|
| 18 |
+
model = AutoModelForCausalLM.from_pretrained("Anurag02/LLM")
|
| 19 |
+
tokenizer = AutoTokenizer.from_pretrained("Anurag02/LLM")
|
| 20 |
+
|
| 21 |
+
# Example usage for distance calculation
|
| 22 |
+
prompt = """What is the distance from Los Angeles to New York? (Based on the function name, the "origin" and "destination" are flipped for the question)"""
|
| 23 |
+
|
| 24 |
+
# Generate response
|
| 25 |
+
inputs = tokenizer(prompt, return_tensors="pt")
|
| 26 |
+
outputs = model.generate(**inputs)
|
| 27 |
+
response = tokenizer.decode(outputs[0])
|
| 28 |
+
```
|
| 29 |
+
|
| 30 |
+
## Parameters
|
| 31 |
+
- Model Size: ≤ 8B parameters
|
| 32 |
+
- Specialized in: Tool calling tasks
|
| 33 |
+
- Optimized for: Distance calculations with parameter flipping
|
| 34 |
+
|
| 35 |
+
## Example Outputs
|
| 36 |
+
|
| 37 |
+
For the query "What is the distance from Los Angeles to New York?":
|
| 38 |
+
```python
|
| 39 |
+
calculate_distance(origin="New York", destination="Los Angeles")
|
| 40 |
+
```
|