Initial GPTQ model commit

README.md

@@ -45,11 +45,12 @@ pip3 install git+https://github.com/huggingface/transformers
 
 If using a UI like text-generation-webui, make sure to do this in the Python environment of text-generation-webui.
 
+Note that at the time of writing, ExLlama is not yet compatible with the Llama 2 70B models, but support is coming soon.
 
 ## Repositories available
 
 * [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Llama-2-70B-chat-GPTQ)
-* [Original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/meta-llama/Llama-2-70b-chat-hf)
+* [Original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/TheBloke/Llama-2-70B-chat-fp16)
 
 ## Prompt template: Llama-2-Chat
 
@@ -67,14 +68,14 @@ Each separate quant is in a different branch.  See below for instructions on fet
 
 | Branch | Bits | Group Size | Act Order (desc_act) | File Size | ExLlama Compatible? | Made With | Description |
 | ------ | ---- | ---------- | -------------------- | --------- | ------------------- | --------- | ----------- |
-| main | 4 | None | True | 35.33 GB | False | AutoGPTQ | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
-| gptq-4bit-32g-actorder_True | 4 | 32 | False | Still processing | True | AutoGPTQ | 4-bit, with Act Order and group size. 32g gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
-| gptq-4bit-64g-actorder_True | 4 | 64 | False | Still processing | True | AutoGPTQ | 4-bit, with Act Order and group size. 64g uses less VRAM than 32g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
-| gptq-4bit-128g-actorder_True | 4 | 128 | False | 36.65 GB | True | AutoGPTQ | 4-bit, with Act Order and group size. 128g uses even less VRAM, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
-| gptq-3bit--1g-actorder_True | 3 | None | False | Still processing | False | AutoGPTQ | 3-bit, with Act Order and no group size. Lowest possible VRAM requirements. May be lower quality than 3-bit 128g. |
-| gptq-3bit-128g-actorder_False | 3 | 128 | False | Still processing | False | AutoGPTQ | 3-bit, with group size 128g but no act-order. Slightly higher VRAM requirements than 3-bit None. |
-| gptq-3bit-128g-actorder_True | 3 | 128 | False | Still processing | False | AutoGPTQ | 3-bit, with group size 128g and act-order. Higher quality than 128g-False but poor AutoGPTQ CUDA speed. |
-| gptq-3bit-64g-actorder_True | 3 | 64 | False | Still processing | False | AutoGPTQ | 3-bit, with group size 64g and act-order. Highest quality 3-bit option. Poor AutoGPTQ CUDA speed. |
+| main | 4 | 128 | False | 35332232264.00 GB | False | AutoGPTQ | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
+| gptq-4bit-32g-actorder_True | 4 | 32 | True | 40.66 GB | False | AutoGPTQ | 4-bit, with Act Order and group size. 32g gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
+| gptq-4bit-64g-actorder_True | 4 | 64 | True | 37.99 GB | False | AutoGPTQ | 4-bit, with Act Order and group size. 64g uses less VRAM than 32g, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
+| gptq-4bit-128g-actorder_True | 4 | 128 | True | 36.65 GB | False | AutoGPTQ | 4-bit, with Act Order and group size. 128g uses even less VRAM, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
+| gptq-3bit--1g-actorder_True | 3 | None | True | 26.78 GB | False | AutoGPTQ | 3-bit, with Act Order and no group size. Lowest possible VRAM requirements. May be lower quality than 3-bit 128g. |
+| gptq-3bit-128g-actorder_False | 3 | 128 | False | 28.03 GB | False | AutoGPTQ | 3-bit, with group size 128g but no act-order. Slightly higher VRAM requirements than 3-bit None. |
+| gptq-3bit-128g-actorder_True | 3 | 128 | True | 28.03 GB | False | AutoGPTQ | 3-bit, with group size 128g and act-order. Higher quality than 128g-False but poor AutoGPTQ CUDA speed. |
+| gptq-3bit-64g-actorder_True | 3 | 64 | True | 29.30 GB | False | AutoGPTQ | 3-bit, with group size 64g and act-order. Highest quality 3-bit option. Poor AutoGPTQ CUDA speed. |
 
 ## How to download from branches
 
@@ -91,38 +92,44 @@ Please make sure you're using the latest version of [text-generation-webui](http
 
 It is strongly recommended to use the text-generation-webui one-click-installers unless you know how to make a manual install.
 
-Remember to update Transformers to the latest Github version:
+Before trying the model, first update Transformers to the latest Github code:
+
 ```
 pip3 install git+https://github.com/huggingface/transformers
 ```
 
-ExLlama is not currently compatible with Llama 2 70B.
+ExLlama is not currently compatible with Llama 2 70B but support is expected soon.
 
 1. Click the **Model tab**.
-2. Under **Download custom model or LoRA**, enter `TheBloke/Llama-2-70B-chat-GPTQ`.
+2. Under **Download custom model or LoRA**, enter `%%REPO_GPTQ`.
   - To download from a specific branch, enter for example `TheBloke/Llama-2-70B-chat-GPTQ:gptq-4bit-32g-actorder_True`
   - see Provided Files above for the list of branches for each option.
 3. Click **Download**.
 4. The model will start downloading. Once it's finished it will say "Done"
-5. In the top left, click the refresh icon next to **Model**.
-6. In the **Model** dropdown, choose the model you just downloaded: `Llama-2-70B-chat-GPTQ`
-7. The model will automatically load, and is now ready for use!
-8. 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.
-  * Note that you do not need to set GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
-9. Once you're ready, click the **Text Generation tab** and enter a prompt to get started!
+5. Set Loader to AutoGPTQ or GPTQ-for-LLaMA
+  - If you use AutoGPTQ, make sure "No inject fused attention" is ticked
+6. In the top left, click the refresh icon next to **Model**.
+7. In the **Model** dropdown, choose the model you just downloaded: `TheBloke/Llama-2-70B-chat-GPTQ`
+8. The model will automatically load, and is now ready for use!
+9. Then click **Save settings for this model** followed by **Reload the Model** in the top right to make sure your settings are persisted.
+10. Once you're ready, click the **Text Generation tab** and enter a prompt to get started!
 
 ## How to use this GPTQ model from Python code
 
 First make sure you have [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ) installed:
 
-`GITHUB_ACTIONS=true pip install auto-gptq`
+```
+GITHUB_ACTIONS=true pip3 install auto-gptq
+```
 
-Also update Transformers to the latest Github version:
+You also need the latest Transformers code from Github:
 
 ```
 pip3 install git+https://github.com/huggingface/transformers
 ```
 
+You must set `inject_fused_attention=False` as shown below.
+
 Then try the following example code:
 
 ```python
@@ -130,16 +137,17 @@ from transformers import AutoTokenizer, pipeline, logging
 from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
 
 model_name_or_path = "TheBloke/Llama-2-70B-chat-GPTQ"
-model_basename = "gptq_model-4bit--1g"
+model_basename = "gptq_model-4bit-128g"
 
 use_triton = False
 
 tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
 
 model = AutoGPTQForCausalLM.from_quantized(model_name_or_path,
-        model_basename=model_basename
+        model_basename=model_basename,
+        inject_fused_attention=False, # Required for Llama 2 70B model at this time.
         use_safetensors=True,
-        trust_remote_code=True,
+        trust_remote_code=False,
         device="cuda:0",
         use_triton=use_triton,
         quantize_config=None)
@@ -150,16 +158,17 @@ To download from a specific branch, use the revision parameter, as in this examp
 model = AutoGPTQForCausalLM.from_quantized(model_name_or_path,
         revision="gptq-4bit-32g-actorder_True",
         model_basename=model_basename,
+        inject_fused_attention=False, # Required for Llama 2 70B model at this time.
         use_safetensors=True,
-        trust_remote_code=True,
+        trust_remote_code=False,
         device="cuda:0",
         quantize_config=None)
 """
 
 prompt = "Tell me about AI"
-prompt_template=f'''System: You are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe.  Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature. If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.
-User: {prompt}
-Assistant:
+prompt_template=f'''SYSTEM: You are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe.  Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature. If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.
+USER: {prompt}
+ASSISTANT:
 '''
 
 print("\n\n*** Generate:")
@@ -191,7 +200,83 @@ print(pipe(prompt_template)[0]['generated_text'])
 
 The files provided will work with AutoGPTQ (CUDA and Triton modes), GPTQ-for-LLaMa (only CUDA has been tested), and Occ4m's GPTQ-for-LLaMa fork.
 
-ExLlama does not currently work with Llama 2 70B models.
+ExLlama is not currently compatible with Llama 2 70B models, but support is coming soon. Please see the Provided Files table above for per-file compatibility.
+
+<!-- footer start -->
+## Discord
+
+For further support, and discussions on these models and AI in general, join us at:
+
+[TheBloke AI's Discord server](https://discord.gg/theblokeai)
+
+## Thanks, and how to contribute.
+
+Thanks to the [chirper.ai](https://chirper.ai) team!
+
+I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
+
+If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
+
+Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
+
+* Patreon: https://patreon.com/TheBlokeAI
+* Ko-Fi: https://ko-fi.com/TheBlokeAI
+
+**Special thanks to**: Luke from CarbonQuill, Aemon Algiz.
+
+**Patreon special mentions**: Space Cruiser, Nikolai Manek, Sam, Chris McCloskey, Rishabh Srivastava, Kalila, Spiking Neurons AB, Khalefa Al-Ahmad, WelcomeToTheClub, Chadd, Lone Striker, Viktor Bowallius, Edmond Seymore, Ai Maven, Chris Smitley, Dave, Alexandros Triantafyllidis, Luke @flexchar, Elle, ya boyyy, Talal Aujan, Alex , Jonathan Leane, Deep Realms, Randy H, subjectnull, Preetika Verma, Joseph William Delisle, Michael Levine, chris gileta, K, Oscar Rangel, LangChain4j, Trenton Dambrowitz, Eugene Pentland, Johann-Peter Hartmann, Femi Adebogun, Illia Dulskyi, senxiiz, Daniel P. Andersen, Sean Connelly, Artur Olbinski, RoA, Mano Prime, Derek Yates, Raven Klaugh, David Flickinger, Willem Michiel, Pieter, Willian Hasse, vamX, Luke Pendergrass, webtim, Ghost , Rainer Wilmers, Nathan LeClaire, Will Dee, Cory Kujawski, John Detwiler, Fred von Graf, biorpg, Iucharbius , Imad Khwaja, Pierre Kircher, terasurfer , Asp the Wyvern, John Villwock, theTransient, zynix , Gabriel Tamborski, Fen Risland, Gabriel Puliatti, Matthew Berman, Pyrater, SuperWojo, Stephen Murray, Karl Bernard, Ajan Kanaga, Greatston Gnanesh, Junyu Yang.
+
+Thank you to all my generous patrons and donaters!
+
+<!-- footer end -->
+
+# Original model card: Meta's Llama 2 70B Chat
+
+
+<!-- header start -->
+<div style="width: 100%;">
+    <img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
+</div>
+<div style="display: flex; justify-content: space-between; width: 100%;">
+    <div style="display: flex; flex-direction: column; align-items: flex-start;">
+        <p><a href="https://discord.gg/theblokeai">Chat & support: my new Discord server</a></p>
+    </div>
+    <div style="display: flex; flex-direction: column; align-items: flex-end;">
+        <p><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
+    </div>
+</div>
+<!-- header end -->
+
+# Meta's Llama 2 70B Chat fp16
+
+These files are fp16 pytorch model files for [Meta's Llama 2 70B Chat](https://huggingface.co/meta-llama/Llama-2-70b-chat-hf).
+
+They were produced by downloading the PTH files from Meta, and then converting to HF format using the latest Transformers 4.32.0.dev0, from Git, with the Llama 2 PR included: https://github.com/huggingface/transformers/pull/24891.
+
+Command to convert was:
+```
+ python3 /workspace/venv/pytorch2/lib/python3.10/site-packages/transformers/models/llama/convert_llama_weights_to_hf.py --input_dir /workspace/git/llama/download --model_size 70B --output_dir /workspace/process/llama-2-70b-chat/source --safe_serialization true
+```
+
+The files were saved in Safetensors format.
+
+I am uploading this repo because I initially tried to create GPTQs using the [Meta Llama 2 70B Chat HF repo](https://huggingface.co/meta-llama/Llama-2-70b-chat-hf), but got strange errors that suggested the weights were not correct.  But converting from the PTH files using the latest `convert_llama_weights_to_hf.py` script worked fine.
+
+Many thanks to William Beauchamp from [Chai](https://chai-research.com/) for providing the hardware for these quantisations!
+
+## Repositories available
+
+* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Llama-2-70B-chat-GPTQ)
+* [Original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/meta-llama/Llama-2-70b-chat-hf)
+* [My fp16 conversion of the unquantised PTH model files](https://huggingface.co/TheBloke/Llama-2-70B-chat-fp16)
+
+## Prompt template: Llama-2-Chat
+
+```
+System: You are a helpful, respectful and honest assistant. Always answer as helpfully as possible, while being safe.  Your answers should not include any harmful, unethical, racist, sexist, toxic, dangerous, or illegal content. Please ensure that your responses are socially unbiased and positive in nature. If a question does not make any sense, or is not factually coherent, explain why instead of answering something not correct. If you don't know the answer to a question, please don't share false information.
+User: {prompt}
+Assistant:
+```
 
 <!-- footer start -->
 ## Discord