zephyr-7b-gemma-sft-african-ultrachat-5k

This model is a fine-tuned version of google/gemma-7b on the masakhane/african-ultrachat and the israel/untrachat_en datasets. It achieves the following results on the evaluation set:

• Loss: 1.1356

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 1e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 8
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• total_eval_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
1.1994	1.0	2480	1.1954
1.0039	2.0	4960	1.0974
0.6836	3.0	7440	1.1356

Framework versions

• Transformers 4.39.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.14.6
• Tokenizers 0.15.2

Usage

# Install transformers from source - only needed for versions <= v4.34
# pip install git+https://github.com/huggingface/transformers.git
# pip install accelerate

import torch
from transformers import pipeline

pipe = pipeline("text-generation", model="
zephyr-7b-gemma-sft-african-ultrachat-5k", torch_dtype=torch.bfloat16, device_map="auto")

# We use the tokenizer's chat template to format each message - see https://huggingface.co/docs/transformers/main/en/chat_templating
messages = [
    {
        "role": "system",
        "content": "You are a friendly chatbot who answewrs question in given language",
    },
    {"role": "user", "content": "what is the 3 biggest countrys in Africa?"},
]
prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
outputs = pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])
# <|system|>
# You are a friendly chatbot who always responds in the style of a pirate<eos>
# <|user|>
# what is the 3 biggest countrys in Africa?<eos>
# <|assistant|>
# The 3 biggest countries in Africa are Nigeria, Ethiopia and South Africa.

Quantized Versions through bitsandbytes

import torch
from transformers import pipeline
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig


quantization_config = BitsAndBytesConfig(load_in_4bit=True)

tokenizer = AutoTokenizer.from_pretrained("
zephyr-7b-gemma-sft-african-ultrachat-5k")
model = AutoModelForCausalLM.from_pretrained("
zephyr-7b-gemma-sft-african-ultrachat-5k", quantization_config=quantization_config)


pipe = pipeline("text-generation", model=model,tokenizer=tokenizer, torch_dtype=torch.bfloat16, device_map="auto")

messages = [
    {
        "role": "system",
        "content": "You are a friendly chatbot who answewrs question in given language",
    },
    {"role": "user", "content": "list languages in Africa?"},
]
prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
outputs = pipe(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])