Update README.md
Browse files
README.md
CHANGED
|
@@ -1 +1,32 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model:
|
| 3 |
+
- unsloth/Llama-3.2-1B
|
| 4 |
+
tags:
|
| 5 |
+
- text-generation-inference
|
| 6 |
+
- transformers
|
| 7 |
+
- math
|
| 8 |
+
- conversational
|
| 9 |
+
- llama
|
| 10 |
+
- meta
|
| 11 |
+
license: apache-2.0
|
| 12 |
+
language:
|
| 13 |
+
- en
|
| 14 |
+
library_name: transformers
|
| 15 |
+
---
|
| 16 |
+
# GsMath-Llama-1B
|
| 17 |
+
## Model Description:
|
| 18 |
+
This is a fine-tuned version of [unsloth/Llama-3.2-1B](https://huggingface.co/unsloth/Llama-3.2-1B)!
|
| 19 |
+
- **recommended settings for inference:** min_p = 0.1 and temperature = 1.5 , Read this [Tweet](https://x.com/menhguin/status/1826132708508213629) to understand why.
|
| 20 |
+
- **License :** apache-2.0
|
| 21 |
+
- **Finetuned from model :** unsloth/Llama-3.2-1B
|
| 22 |
+
## Benchmarks:
|
| 23 |
+
We evaluate both models on GSM8K using the standard lm-eval 5-shot exact-match protocol. Under identical decoding and extraction settings,GsMath-Llama-1B outperforms Meta’s Llama-3.2-1B by 2x,demonstrating an improvement in small-model mathematical capability.
|
| 24 |
+
| Model | Params | GSM8K (5-shot, EM) |
|
| 25 |
+
| ----------------------------- | ------ | ------------------ |
|
| 26 |
+
| **GsMath-Llama-1B** | 1B | **13.7%** |
|
| 27 |
+
| Llama-3.2-1B | 1B | 6.8% |
|
| 28 |
|
| 29 |
+
|
| 30 |
+
<p align="center">
|
| 31 |
+
<img alt="GsMath-Llama-1B" src="https://huggingface.co/Cannae-AI/ReasoningLlama-Math-1B-IT/resolve/main/ChatGPT%20Image%2018%20nov.%202025%2C%2020_55_23.png">
|
| 32 |
+
</p>
|