duplicated_from: localmodels/LLM

Wizard Vicuna 13B Uncensored GPTQ

From: https://huggingface.co/ehartford/Wizard-Vicuna-13B-Uncensored merged with SuperHOT 8K.

This is an experimental new GPTQ which offers up to 8K context size

The increased context is tested to work with ExLlama, via the latest release of text-generation-webui.

It has also been tested from Python code using AutoGPTQ and trust_remote_code=True.

Please read carefully below to see how to use it.

How to use this model in text-generation-webui with ExLlama

Using the latest version of text-generation-webui:

1. Click the Model tab.
2. Under Download custom model or LoRA, enter localmodels/Wizard-Vicuna-13B-Uncensored-SuperHOT-8K-GPTQ.
3. Click Download.
4. The model will start downloading. Once it's finished it will say "Done"
5. Untick Autoload the model
6. In the top left, click the refresh icon next to Model.
7. In the Model dropdown, choose the model you just downloaded: Wizard-Vicuna-13B-Uncensored-SuperHOT-8K-GPTQ
8. To use the increased context, set the Loader to ExLlama, set max_seq_len to 8192 or 4096, and set compress_pos_emb to 4 for 8192 context, or to 2 for 4096 context.
9. Now click Save Settings followed by Reload
10. The model will automatically load, and is now ready for use!
11. 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 with AutoGPTQ

First make sure you have AutoGPTQ and Einops installed:

pip3 install einops auto-gptq

Then run the following code. Note that in order to get this to work, config.json has been hardcoded to a sequence length of 8192.

If you want to try 4096 instead to reduce VRAM usage, manually edit config.json to set max_position_embeddings to the value you want.

from transformers import AutoTokenizer, pipeline, logging
from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
import argparse

model_name_or_path = "TheBloke/Wizard-Vicuna-13B-Uncensored-SuperHOT-8K-GPTQ"
model_basename = "wizard-vicuna-13b-uncensored-superhot-8k-GPTQ-4bit-128g.no-act.order"

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,
        use_safetensors=True,
        trust_remote_code=True,
        device_map='auto',
        use_triton=use_triton,
        quantize_config=None)

model.seqlen = 8192

# Note: check the prompt template is correct for this model.
prompt = "Tell me about AI"
prompt_template=f'''USER: {prompt}
ASSISTANT:'''

print("\n\n*** Generate:")

input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))

# Inference can also be done using transformers' pipeline

# Prevent printing spurious transformers error when using pipeline with AutoGPTQ
logging.set_verbosity(logging.CRITICAL)

print("*** Pipeline:")
pipe = pipeline(
    "text-generation",
    model=model,
    tokenizer=tokenizer,
    max_new_tokens=512,
    temperature=0.7,
    top_p=0.95,
    repetition_penalty=1.15
)

print(pipe(prompt_template)[0]['generated_text'])

Model

wizard-vicuna-13b-uncensored-superhot-8k-GPTQ-4bit-128g.no-act.order.safetensors

This will work with AutoGPTQ, ExLlama, 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.

It was created with group_size 128 to increase inference accuracy, but without --act-order (desc_act) to increase compatibility and improve inference speed.

• wizard-vicuna-13b-uncensored-superhot-8k-GPTQ-4bit-128g.no-act.order.safetensors
  • Works for use with ExLlama with increased context (4096 or 8192)
  • Works with AutoGPTQ in Python code, including with increased context if trust_remote_code=True is set.
  • Parameters: Groupsize = 128. No act-order.

SuperHOT Prototype 2 w/ 8K Context

This is a second prototype of SuperHOT, this time 30B with 8K context and no RLHF, using the same technique described in the github blog.
Tests have shown that the model does indeed leverage the extended context at 8K.

Looking for Merged & Quantized Models?

• 30B 4-bit CUDA: tmpupload/superhot-30b-8k-4bit-safetensors
• 30B 4-bit CUDA 128g: tmpupload/superhot-30b-8k-4bit-128g-safetensors

Training Details

I trained the LoRA with the following configuration:

• 1200 samples (~400 samples over 2048 sequence length)
• learning rate of 3e-4
• 3 epochs
• The exported modules are:
  • q_proj
  • k_proj
  • v_proj
  • o_proj
  • no bias
• Rank = 4
• Alpha = 8
• no dropout
• weight decay of 0.1
• AdamW beta1 of 0.9 and beta2 0.99, epsilon of 1e-5
• Trained on 4-bit base model