Update README.md
Browse files
README.md
CHANGED
@@ -1,6 +1,13 @@
|
|
1 |
---
|
2 |
inference: false
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
---
|
5 |
|
6 |
<!-- header start -->
|
@@ -26,7 +33,7 @@ It is the result of quantising to 4bit using [GPTQ-for-LLaMa](https://github.com
|
|
26 |
## Repositories available
|
27 |
|
28 |
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/WizardCoder-Guanaco-15B-V1.0-GPTQ)
|
29 |
-
* [
|
30 |
* [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/LoupGarou/WizardCoder-Guanaco-15B-V1.0)
|
31 |
|
32 |
## Prompt template: Alpaca
|
@@ -53,9 +60,10 @@ It is strongly recommended to use the text-generation-webui one-click-installers
|
|
53 |
5. In the top left, click the refresh icon next to **Model**.
|
54 |
6. In the **Model** dropdown, choose the model you just downloaded: `WizardCoder-Guanaco-15B-V1.0-GPTQ`
|
55 |
7. The model will automatically load, and is now ready for use!
|
56 |
-
8. If you
|
|
|
57 |
* Note that you do not need to set GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
|
58 |
-
|
59 |
|
60 |
## How to use this GPTQ model from Python code
|
61 |
|
@@ -127,14 +135,14 @@ print(pipe(prompt_template)[0]['generated_text'])
|
|
127 |
|
128 |
This will work with AutoGPTQ and CUDA versions of GPTQ-for-LLaMa. There are reports of issues with Triton mode of recent GPTQ-for-LLaMa. If you have issues, please use AutoGPTQ instead.
|
129 |
|
130 |
-
|
131 |
|
132 |
It was created with group_size 128 to increase inference accuracy, but without --act-order (desc_act) to increase compatibility and improve inference speed.
|
133 |
|
134 |
* `wizardcoder-guanaco-15b-v1.0-GPTQ-4bit-128g.no-act.order.safetensors`
|
135 |
* Works with AutoGPTQ in CUDA or Triton modes.
|
136 |
-
* [ExLlama](https://github.com/turboderp/exllama)
|
137 |
-
*
|
138 |
* Works with text-generation-webui, including one-click-installers.
|
139 |
* Parameters: Groupsize = 128. Act Order / desc_act = False.
|
140 |
|
|
|
1 |
---
|
2 |
inference: false
|
3 |
+
language:
|
4 |
+
- en
|
5 |
+
datasets:
|
6 |
+
- guanaco
|
7 |
+
model_hub_library:
|
8 |
+
- transformers
|
9 |
+
license:
|
10 |
+
- apache-2.0
|
11 |
---
|
12 |
|
13 |
<!-- header start -->
|
|
|
33 |
## Repositories available
|
34 |
|
35 |
* [4-bit GPTQ models for GPU inference](https://huggingface.co/TheBloke/WizardCoder-Guanaco-15B-V1.0-GPTQ)
|
36 |
+
* [4, 5, and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/WizardCoder-Guanaco-15B-V1.0-GGML)
|
37 |
* [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/LoupGarou/WizardCoder-Guanaco-15B-V1.0)
|
38 |
|
39 |
## Prompt template: Alpaca
|
|
|
60 |
5. In the top left, click the refresh icon next to **Model**.
|
61 |
6. In the **Model** dropdown, choose the model you just downloaded: `WizardCoder-Guanaco-15B-V1.0-GPTQ`
|
62 |
7. The model will automatically load, and is now ready for use!
|
63 |
+
8. If you have problems, make sure that **Loader** is set to **AutoGPTQ**.
|
64 |
+
9. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
|
65 |
* Note that you do not need to set GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
|
66 |
+
10. Once you're ready, click the **Text Generation tab** and enter a prompt to get started!
|
67 |
|
68 |
## How to use this GPTQ model from Python code
|
69 |
|
|
|
135 |
|
136 |
This will work with AutoGPTQ and CUDA versions of GPTQ-for-LLaMa. There are reports of issues with Triton mode of recent GPTQ-for-LLaMa. If you have issues, please use AutoGPTQ instead.
|
137 |
|
138 |
+
As this is not a Llama model, it will not be supported by ExLlama.
|
139 |
|
140 |
It was created with group_size 128 to increase inference accuracy, but without --act-order (desc_act) to increase compatibility and improve inference speed.
|
141 |
|
142 |
* `wizardcoder-guanaco-15b-v1.0-GPTQ-4bit-128g.no-act.order.safetensors`
|
143 |
* Works with AutoGPTQ in CUDA or Triton modes.
|
144 |
+
* Does NOT work with [ExLlama](https://github.com/turboderp/exllama).
|
145 |
+
* Untested with GPTQ-for-LLaMa.
|
146 |
* Works with text-generation-webui, including one-click-installers.
|
147 |
* Parameters: Groupsize = 128. Act Order / desc_act = False.
|
148 |
|