Update README.md
Browse files
README.md
CHANGED
@@ -21,11 +21,20 @@ To use, pass `trust_remote_code=True` when loading the model, for example
|
|
21 |
|
22 |
```python
|
23 |
model = AutoModelForCausalLM.from_pretrained("NousResearch/Yarn-Mistral-7b-64k",
|
|
|
24 |
torch_dtype=torch.bfloat16,
|
25 |
device_map="auto",
|
26 |
trust_remote_code=True)
|
27 |
```
|
28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
29 |
## Collaborators
|
30 |
|
31 |
- [bloc97](https://github.com/bloc97): Methods, paper and evals
|
|
|
21 |
|
22 |
```python
|
23 |
model = AutoModelForCausalLM.from_pretrained("NousResearch/Yarn-Mistral-7b-64k",
|
24 |
+
use_flash_attention_2=True,
|
25 |
torch_dtype=torch.bfloat16,
|
26 |
device_map="auto",
|
27 |
trust_remote_code=True)
|
28 |
```
|
29 |
|
30 |
+
## Benchmarks
|
31 |
+
|
32 |
+
| Model | Context Window | ARC-c | Hellaswag | MMLU | Truthful QA |
|
33 |
+
|-------|---------------:|------:|----------:|-----:|------------:|
|
34 |
+
| [Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) | 8K | 59.98 | 83.31 | 64.16 | 42.15 |
|
35 |
+
| [Yarn-Mistral-7b-64k](https://huggingface.co/NousResearch/Yarn-Mistral-7b-64k) | 64K | 59.38 | 81.21 | 61.32 | 42.50 |
|
36 |
+
| [Yarn-Mistral-7b-128k](https://huggingface.co/NousResearch/Yarn-Mistral-7b-128k) | 128K | 58.87 | 80.58 | 60.64 | 42.46 |
|
37 |
+
|
38 |
## Collaborators
|
39 |
|
40 |
- [bloc97](https://github.com/bloc97): Methods, paper and evals
|