|
--- |
|
license: mit |
|
datasets: |
|
- starmpcc/Asclepius-Synthetic-Clinical-Notes |
|
language: |
|
- en |
|
--- |
|
## Overview |
|
|
|
This model, elucidator8918/apigen-prototype-0.1, is tailored for API generation, based on the Mistral-7B-Instruct-v0.1-sharded architecture fine-tuned on the LLAMA-2 Instruct 121k Code dataset. |
|
|
|
## Key Information |
|
|
|
- **Model Name**: Mistral-7B-Instruct-v0.1-sharded |
|
- **Fine-tuned Model Name**: elucidator8918/apigen-prototype-0.1 |
|
- **Dataset**: emre/llama-2-instruct-121k-code |
|
- **Language**: English (en) |
|
|
|
## Model Details |
|
|
|
- **LoRA Parameters (QLoRA):** |
|
- LoRA attention dimension: 64 |
|
- Alpha parameter for LoRA scaling: 16 |
|
- Dropout probability for LoRA layers: 0.1 |
|
|
|
- **bitsandbytes Parameters:** |
|
- Activate 4-bit precision base model loading |
|
- Compute dtype for 4-bit base models: float16 |
|
- Quantization type: nf4 |
|
- Activate nested quantization for 4-bit base models: No |
|
|
|
- **TrainingArguments Parameters:** |
|
- Number of training epochs: 1 |
|
- Batch size per GPU for training: 4 |
|
- Batch size per GPU for evaluation: 4 |
|
- Gradient accumulation steps: 1 |
|
- Enable gradient checkpointing: Yes |
|
- Maximum gradient norm: 0.3 |
|
- Initial learning rate: 2e-4 |
|
- Weight decay: 0.001 |
|
- Optimizer: paged_adamw_32bit |
|
- Learning rate scheduler type: cosine |
|
- Warm-up ratio: 0.03 |
|
- Group sequences into batches with the same length: Yes |
|
|
|
## Usage |
|
|
|
|
|
- **Example Code (API Generation):** |
|
|
|
```python |
|
from transformers import pipeline |
|
|
|
api_gen_pipeline = pipeline("text-generation", model="elucidator8918/apigen-prototype-0.1") |
|
|
|
# Run text generation pipeline with our next model |
|
prompt = "Write code to do a POST request in FastAPI framework to find the multiplication of two matrices using NumPy" |
|
pipe = pipeline(task="text-generation", model=model, tokenizer=tokenizer, max_length=500) |
|
result = pipe(f"[INST] {prompt} [/INST]") |
|
print(result[0]['generated_text']) |
|
``` |
|
|
|
- **Output API Generation:** |
|
``` |
|
[INST] Write code to do a POST request in fastapi framework to find the multiplication of two matrices using numpy [/INST] |
|
Below is an example of how to make a POST request in FastAPI to find the multiplication of two matrices using numpy: |
|
``` |
|
```python |
|
from fastapi import FastAPI, HTTPException |
|
import numpy as np |
|
|
|
app = FastAPI() |
|
|
|
@app.post("/matrix_multiplication") |
|
async def matrix_multiplication(matrix1: np.ndarray, matrix2: np.ndarray): |
|
if matrix1.shape[1]!= matrix2.shape[0]: |
|
raise HTTPException(status_code=400, detail="The number of columns in matrix1 must be equal to the number of rows in matrix2") |
|
result = np.matmul(matrix1, matrix2) |
|
return {"result": result} |
|
``` |
|
This code defines a FastAPI endpoint at `/matrix_multiplication` that takes two matrices as input and returns the multiplication of the two matrices. The `np.matmul` function is used to perform the multiplication. The endpoint also includes a check to ensure that the number of columns in the first matrix is equal to the number of rows in the second matrix. |
|
|
|
To use this endpoint, you can make a POST request to `http://localhost:8000/matrix_multiplication` with the two matrices as input. The response will include the multiplication of the two matrices. |
|
```python |
|
import requests |
|
|
|
matrix1 = np.array([[1, 2], [3, 4]]) |
|
matrix2 = np.array([[5, 6], [7, 8]]) |
|
|
|
response = requests.post("http://localhost:8000/matrix_multiplication", json={"matrix1": matrix1, "matrix2": matrix2}) |
|
|
|
print(response.json()) |
|
``` |
|
This code makes a POST request to the endpoint with the two matrices as input and prints the response. The response should include the multiplication of the two matrices, which is `[[11, 14], [29, 36]]`. |
|
|
|
## License |
|
|
|
This model is released under the MIT License. |