31 lines
926 B
Markdown
31 lines
926 B
Markdown
|
---
|
||
|
|
license: apache-2.0
|
||
|
|
base_model: Qwen/Qwen2.5-Coder-1.5B-Instruct
|
||
|
|
tags:
|
||
|
|
- code
|
||
|
|
- qlora
|
||
|
|
- qwen2.5
|
||
|
|
library_name: transformers
|
||
|
|
---
|
||
|
|
|
||
|
|
# Index-9
|
||
|
|
|
||
|
|
`Qwen2.5-Coder-1.5B-Instruct` fine-tuned with QLoRA for coding (Python + web:
|
||
|
|
JS/TS/HTML/CSS). Designed to be paired with live **web search** — a small model
|
||
|
|
that leans on the web for facts. See the companion Space for the search-augmented
|
||
|
|
chat demo.
|
||
|
|
|
||
|
|
## Usage
|
||
|
|
|
||
|
|
```python
|
||
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
||
|
|
|
||
|
|
tok = AutoTokenizer.from_pretrained("Visixn/Index-9")
|
||
|
|
model = AutoModelForCausalLM.from_pretrained("Visixn/Index-9")
|
||
|
|
|
||
|
|
msgs = [{"role": "user", "content": "Write a palindrome checker in Python"}]
|
||
|
|
prompt = tok.apply_chat_template(msgs, tokenize=False, add_generation_prompt=True)
|
||
|
|
out = model.generate(**tok(prompt, return_tensors="pt"), max_new_tokens=256)
|
||
|
|
print(tok.decode(out[0], skip_special_tokens=True))
|
||
|
|
```
|