Update README.md
Browse files
README.md
CHANGED
|
@@ -40,4 +40,8 @@ input = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_p
|
|
| 40 |
encoded_input = tokenizer(input, return_tensors = "pt").to(model.device)
|
| 41 |
output = model.generate(**encoded_input, max_new_tokens=1024)
|
| 42 |
print(tokenizer.decode(output[0], skip_special_tokens=False))
|
| 43 |
-
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
encoded_input = tokenizer(input, return_tensors = "pt").to(model.device)
|
| 41 |
output = model.generate(**encoded_input, max_new_tokens=1024)
|
| 42 |
print(tokenizer.decode(output[0], skip_special_tokens=False))
|
| 43 |
+
```
|
| 44 |
+
|
| 45 |
+
### Evaluation
|
| 46 |
+
The model was evaluated on a randomly sampled subset of 1,000 records from the test split of the [Math-QA](https://huggingface.co/datasets/rvv-karma/Math-QA) dataset.
|
| 47 |
+
Math Genius 7B achieved an accuracy of 93.1% in producing the correct final answer under the pass@1 evaluation metric.
|