Hes
#7
by
mmahlwy3
- opened
- README.md +3 -3
- config.json +2 -2
README.md
CHANGED
@@ -27,13 +27,13 @@ The primary intended users of the model are researchers and hobbyists in natural
|
|
27 |
|
28 |
## How to Get Started with the Model
|
29 |
|
30 |
-
|
31 |
-
|
32 |
|
33 |
## Training Details
|
34 |
|
35 |
Vicuna v1.3 is fine-tuned from LLaMA with supervised instruction fine-tuning.
|
36 |
-
The training data is around
|
37 |
See more details in the "Training Details of Vicuna Models" section in the appendix of this [paper](https://arxiv.org/pdf/2306.05685.pdf).
|
38 |
|
39 |
## Evaluation
|
|
|
27 |
|
28 |
## How to Get Started with the Model
|
29 |
|
30 |
+
Command line interface: https://github.com/lm-sys/FastChat#vicuna-weights.
|
31 |
+
APIs (OpenAI API, Huggingface API): https://github.com/lm-sys/FastChat/tree/main#api.
|
32 |
|
33 |
## Training Details
|
34 |
|
35 |
Vicuna v1.3 is fine-tuned from LLaMA with supervised instruction fine-tuning.
|
36 |
+
The training data is around 140K conversations collected from ShareGPT.com.
|
37 |
See more details in the "Training Details of Vicuna Models" section in the appendix of this [paper](https://arxiv.org/pdf/2306.05685.pdf).
|
38 |
|
39 |
## Evaluation
|
config.json
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "/home/ubuntu/model_weights/vicuna-33b-
|
3 |
"architectures": [
|
4 |
"LlamaForCausalLM"
|
5 |
],
|
@@ -18,6 +18,6 @@
|
|
18 |
"tie_word_embeddings": false,
|
19 |
"torch_dtype": "float16",
|
20 |
"transformers_version": "4.28.1",
|
21 |
-
"use_cache":
|
22 |
"vocab_size": 32000
|
23 |
}
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "/home/ubuntu/model_weights/vicuna-33b-gpt4-20230521/",
|
3 |
"architectures": [
|
4 |
"LlamaForCausalLM"
|
5 |
],
|
|
|
18 |
"tie_word_embeddings": false,
|
19 |
"torch_dtype": "float16",
|
20 |
"transformers_version": "4.28.1",
|
21 |
+
"use_cache": false,
|
22 |
"vocab_size": 32000
|
23 |
}
|