File size: 2,527 Bytes
9fc3eb3 9614041 9fc3eb3 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 |
---
base_model: unsloth/Qwen2.5-Math-1.5B-Instruct-bnb-4bit
language:
- en
license: apache-2.0
tags:
- text-generation-inference
- transformers
- unsloth
- qwen2
- trl
---
# Uploaded model
- **Developed by:** thanhkt
- **License:** apache-2.0
- **Finetuned from model :** unsloth/Qwen2.5-Math-1.5B-Instruct-bnb-4bit
This qwen2 model was trained 2x faster with [Unsloth](https://github.com/unslothai/unsloth) and Huggingface's TRL library.
## Dataset
The model was trained on the Nvidia-mathinstuct dataset, which consists of 100,000 rows. This dataset was specifically chosen to enhance the model's mathematical reasoning and instruction-following capabilities.
### 🤗 Hugging Face Transformers
Qwen2.5-Math can be deployed and infered in the same way as [Qwen2.5](https://github.com/QwenLM/Qwen2.5). Here we show a code snippet to show you how to use the chat model with `transformers`:
```python
from unsloth import FastLanguageModel
import torch
max_seq_length = 4096 # Choose any! We auto support RoPE Scaling internally!
dtype = None # None for auto detection. Float16 for Tesla T4, V100, Bfloat16 for Ampere+
load_in_4bit = True # Use 4bit quantization to reduce memory usage. Can be False.
model, tokenizer = FastLanguageModel.from_pretrained(
model_name = "thanhkt/Qwen2.5-1.5B-MathInstruct",
max_seq_length = max_seq_length,
dtype = dtype,
load_in_4bit = load_in_4bit,
# token = "hf_...", # use one if using gated models like meta-llama/Llama-2-7b-hf
)
alpaca_prompt = """Below...
### Instruct:
{}
### Input:
{}
### Output:
{}"""
FastLanguageModel.for_inference(model) # Enable native 2x faster inference
inputs = tokenizer(
[
alpaca_prompt.format(
"""A company wants to make a pipeline from a point A on shore to a point B on an island. The island is 6km from the coast. The price to build an onshore pipeline is $50,000 per kilometer, and $130,000 per kilometer to build
underwater. B' is the point on the coast so that BB' is perpendicular to the coast. The distance from A to B' is 9km. Position C on section AB' so that when connecting pipes according to ACB, the amount is minimal. At that time, C is one paragraph away from A by:
A. 6.5km B. 6km C. 0km D.9km""", # instruction
"", # input
"", # output - leave this blank for generation!
)
], return_tensors = "pt").to("cuda")
from transformers import TextStreamer
text_streamer = TextStreamer(tokenizer)
_ = model.generate(**inputs, streamer = text_streamer, max_new_tokens = 512)
```
|