Updating model files
Browse files
README.md
CHANGED
|
@@ -1,6 +1,17 @@
|
|
| 1 |
---
|
| 2 |
license: other
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
# Vicuna 7B 1.1 HF
|
| 5 |
|
| 6 |
This is an HF version of the [Vicuna 7B 1.1 model](https://huggingface.co/lmsys/vicuna-7b-delta-v1.1).
|
|
@@ -15,7 +26,7 @@ I have the following Vicuna 1.1 repositories available:
|
|
| 15 |
* [Unquantized 13B 1.1 model for GPU - HF format](https://huggingface.co/TheBloke/vicuna-13B-1.1-HF)
|
| 16 |
* [GPTQ quantized 4bit 13B 1.1 for GPU - `safetensors` and `pt` formats](https://huggingface.co/TheBloke/vicuna-13B-1.1-GPTQ-4bit-128g)
|
| 17 |
* [GPTQ quantized 4bit 13B 1.1 for CPU - GGML format for `llama.cpp`](https://huggingface.co/TheBloke/vicuna-13B-1.1-GPTQ-4bit-128g-GGML)
|
| 18 |
-
|
| 19 |
**7B models:**
|
| 20 |
* [Unquantized 7B 1.1 model for GPU - HF format](https://huggingface.co/TheBloke/vicuna-7B-1.1-HF)
|
| 21 |
* [GPTQ quantized 4bit 7B 1.1 for GPU - `safetensors` and `pt` formats](https://huggingface.co/TheBloke/vicuna-7B-1.1-GPTQ-4bit-128g)
|
|
@@ -59,4 +70,4 @@ A preliminary evaluation of the model quality is conducted by creating a set of
|
|
| 59 |
|
| 60 |
## Major updates of weights v1.1
|
| 61 |
- Refactor the tokenization and separator. In Vicuna v1.1, the separator has been changed from `"###"` to the EOS token `"</s>"`. This change makes it easier to determine the generation stop criteria and enables better compatibility with other libraries.
|
| 62 |
-
- Fix the supervised fine-tuning loss computation for better model quality.
|
|
|
|
| 1 |
---
|
| 2 |
license: other
|
| 3 |
---
|
| 4 |
+
<div style="width: 100%;">
|
| 5 |
+
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
|
| 6 |
+
</div>
|
| 7 |
+
<div style="display: flex; justify-content: space-between; width: 100%;">
|
| 8 |
+
<div style="display: flex; flex-direction: column; align-items: flex-start;">
|
| 9 |
+
<p><a href="https://discord.gg/UBgz4VXf">Chat & support: my new Discord server</a></p>
|
| 10 |
+
</div>
|
| 11 |
+
<div style="display: flex; flex-direction: column; align-items: flex-end;">
|
| 12 |
+
<p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? Patreon coming soon!</a></p>
|
| 13 |
+
</div>
|
| 14 |
+
</div>
|
| 15 |
# Vicuna 7B 1.1 HF
|
| 16 |
|
| 17 |
This is an HF version of the [Vicuna 7B 1.1 model](https://huggingface.co/lmsys/vicuna-7b-delta-v1.1).
|
|
|
|
| 26 |
* [Unquantized 13B 1.1 model for GPU - HF format](https://huggingface.co/TheBloke/vicuna-13B-1.1-HF)
|
| 27 |
* [GPTQ quantized 4bit 13B 1.1 for GPU - `safetensors` and `pt` formats](https://huggingface.co/TheBloke/vicuna-13B-1.1-GPTQ-4bit-128g)
|
| 28 |
* [GPTQ quantized 4bit 13B 1.1 for CPU - GGML format for `llama.cpp`](https://huggingface.co/TheBloke/vicuna-13B-1.1-GPTQ-4bit-128g-GGML)
|
| 29 |
+
|
| 30 |
**7B models:**
|
| 31 |
* [Unquantized 7B 1.1 model for GPU - HF format](https://huggingface.co/TheBloke/vicuna-7B-1.1-HF)
|
| 32 |
* [GPTQ quantized 4bit 7B 1.1 for GPU - `safetensors` and `pt` formats](https://huggingface.co/TheBloke/vicuna-7B-1.1-GPTQ-4bit-128g)
|
|
|
|
| 70 |
|
| 71 |
## Major updates of weights v1.1
|
| 72 |
- Refactor the tokenization and separator. In Vicuna v1.1, the separator has been changed from `"###"` to the EOS token `"</s>"`. This change makes it easier to determine the generation stop criteria and enables better compatibility with other libraries.
|
| 73 |
+
- Fix the supervised fine-tuning loss computation for better model quality.
|