Update README.md
Browse files
README.md
CHANGED
@@ -20,20 +20,20 @@ We hope that this can enable everyone to finetune their own version of [LLaMA-2-
|
|
20 |
|
21 |
LLaMA-2-7B-32K-Chat is fine-tuned over a combination of two parts:
|
22 |
1. **19K single- and multi-round conversations generated by human instructions and [Llama-2-70B-Chat](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) outputs**.
|
23 |
-
We collected the dataset following the distillation paradigm that is used by Alpaca, Vicuna, WizardLM, Orca — producing instructions by querying a powerful LLM.
|
24 |
-
|
25 |
-
We also share the complete collection
|
26 |
|
27 |
3. **4K instructions of summarization from the BookSum datasets**.
|
28 |
BookSum is a unique dataset designed to address the challenges of long-form narrative summarization.
|
29 |
This dataset features source documents from the literature domain, including novels, plays, and stories, and offers human-written, highly abstractive summaries.
|
30 |
We here focus on chapter-level data. BookSum poses a unique set of challenges, necessitating that the model comprehensively read through each chapter.
|
|
|
31 |
|
32 |
|
33 |
## Model Usage
|
34 |
|
35 |
-
We encourage you to try out this model using the [Together API](https://together.ai/blog/api-announcement).
|
36 |
-
The updated inference stack allows for efficient inference.
|
37 |
Alternatively, you can load the model directly from the Hugging Face model hub using
|
38 |
|
39 |
```python
|
@@ -43,7 +43,7 @@ tokenizer = AutoTokenizer.from_pretrained("togethercomputer/LLaMA-2-7B-32K-Chat"
|
|
43 |
model = AutoModelForCausalLM.from_pretrained("togethercomputer/LLaMA-2-7B-32K-Chat", trust_remote_code=True, torch_dtype=torch.float16)
|
44 |
```
|
45 |
|
46 |
-
The model is also hosted on [Together Playground](https://api.together.xyz/playground). You can simply play with the model by using prompt formatted by
|
47 |
|
48 |
```
|
49 |
[INST] <your instruction here> [\INST].
|
|
|
20 |
|
21 |
LLaMA-2-7B-32K-Chat is fine-tuned over a combination of two parts:
|
22 |
1. **19K single- and multi-round conversations generated by human instructions and [Llama-2-70B-Chat](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf) outputs**.
|
23 |
+
We collected the dataset following the distillation paradigm that is used by Alpaca, Vicuna, WizardLM, Orca — producing instructions by querying a powerful LLM (in this case, [Llama-2-70B-Chat](https://huggingface.co/meta-llama/Llama-2-7b-chat-hf)).
|
24 |
+
The complete dataset is also released [here](https://huggingface.co/datasets/togethercomputer/llama-instruct).
|
25 |
+
We also share the complete recipe for the data collection process [here](https://github.com/togethercomputer/LLaMA-2-32K-Chat).
|
26 |
|
27 |
3. **4K instructions of summarization from the BookSum datasets**.
|
28 |
BookSum is a unique dataset designed to address the challenges of long-form narrative summarization.
|
29 |
This dataset features source documents from the literature domain, including novels, plays, and stories, and offers human-written, highly abstractive summaries.
|
30 |
We here focus on chapter-level data. BookSum poses a unique set of challenges, necessitating that the model comprehensively read through each chapter.
|
31 |
+
We used 4K of the instructions in our fine-tuning.
|
32 |
|
33 |
|
34 |
## Model Usage
|
35 |
|
36 |
+
We encourage you to try out this model using the [Together API](https://together.ai/blog/api-announcement). The updated inference stack allows for efficient inference.
|
|
|
37 |
Alternatively, you can load the model directly from the Hugging Face model hub using
|
38 |
|
39 |
```python
|
|
|
43 |
model = AutoModelForCausalLM.from_pretrained("togethercomputer/LLaMA-2-7B-32K-Chat", trust_remote_code=True, torch_dtype=torch.float16)
|
44 |
```
|
45 |
|
46 |
+
The model is also hosted on [Together Playground](https://api.together.xyz/playground). You can simply play with the model by using prompt formatted by:
|
47 |
|
48 |
```
|
49 |
[INST] <your instruction here> [\INST].
|