Commit
•
fd8e2d7
1
Parent(s):
7beb3b0
Update README.md
Browse files
README.md
CHANGED
@@ -82,7 +82,11 @@ model-index:
|
|
82 |
`stable-code-instruct-3b` is a 2.7B billion parameter decoder-only language model tuned from [`stable-code-3b`](https://huggingface.co/stabilityai/stable-code-3b/). This model was trained on a mix of publicly available datasets, synthetic datasets using [Direct Preference Optimization (DPO)](https://arxiv.org/abs/2305.18290).
|
83 |
|
84 |
This instruct tune demonstrates state-of-the-art performance (compared to models of similar size) on the MultiPL-E metrics across multiple programming languages tested using [BigCode's Evaluation Harness](https://github.com/bigcode-project/bigcode-evaluation-harness/tree/main), and on the code portions of
|
85 |
-
[MT Bench](https://klu.ai/glossary/mt-bench-eval)
|
|
|
|
|
|
|
|
|
86 |
|
87 |
|
88 |
## Usage
|
@@ -152,8 +156,8 @@ output = tokenizer.batch_decode(tokens[:, inputs.input_ids.shape[-1]:], skip_spe
|
|
152 |
| DeepSeek Coder | 1.3B | 4.6 |
|
153 |
| Stable Code Instruct (DPO) | 3B | **5.8**(ours) |
|
154 |
| Stable Code Instruct (SFT) | 3B | 5.5 |
|
155 |
-
| CodeLlama Instruct | 7B | 3.55 |
|
156 |
| DeepSeek Coder | 6.7B | **6.9** |
|
|
|
157 |
| StarChat2 | 15B | 5.7 |
|
158 |
|
159 |
|
|
|
82 |
`stable-code-instruct-3b` is a 2.7B billion parameter decoder-only language model tuned from [`stable-code-3b`](https://huggingface.co/stabilityai/stable-code-3b/). This model was trained on a mix of publicly available datasets, synthetic datasets using [Direct Preference Optimization (DPO)](https://arxiv.org/abs/2305.18290).
|
83 |
|
84 |
This instruct tune demonstrates state-of-the-art performance (compared to models of similar size) on the MultiPL-E metrics across multiple programming languages tested using [BigCode's Evaluation Harness](https://github.com/bigcode-project/bigcode-evaluation-harness/tree/main), and on the code portions of
|
85 |
+
[MT Bench](https://klu.ai/glossary/mt-bench-eval).
|
86 |
+
The model is finetuned to make it useable in tasks like,
|
87 |
+
- General purpose Code/Software Engineering like conversations.
|
88 |
+
- Function Calling
|
89 |
+
- SQL related generation and conversation.
|
90 |
|
91 |
|
92 |
## Usage
|
|
|
156 |
| DeepSeek Coder | 1.3B | 4.6 |
|
157 |
| Stable Code Instruct (DPO) | 3B | **5.8**(ours) |
|
158 |
| Stable Code Instruct (SFT) | 3B | 5.5 |
|
|
|
159 |
| DeepSeek Coder | 6.7B | **6.9** |
|
160 |
+
| CodeLlama Instruct | 7B | 3.55 |
|
161 |
| StarChat2 | 15B | 5.7 |
|
162 |
|
163 |
|