Instructions to use igorktech/RuBit-LLama-63M with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use igorktech/RuBit-LLama-63M with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="igorktech/RuBit-LLama-63M")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("igorktech/RuBit-LLama-63M")
model = AutoModelForCausalLM.from_pretrained("igorktech/RuBit-LLama-63M", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use igorktech/RuBit-LLama-63M with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "igorktech/RuBit-LLama-63M"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "igorktech/RuBit-LLama-63M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/igorktech/RuBit-LLama-63M

SGLang

How to use igorktech/RuBit-LLama-63M with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "igorktech/RuBit-LLama-63M" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "igorktech/RuBit-LLama-63M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "igorktech/RuBit-LLama-63M" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "igorktech/RuBit-LLama-63M",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use igorktech/RuBit-LLama-63M with Docker Model Runner:
```
docker model run hf.co/igorktech/RuBit-LLama-63M
```

RuBit-LLama-63M / README.md

igorktech

Update README.md

9eb6b38 verified over 1 year ago

preview code

Raw

History Blame Contribute Delete

3.23 kB

metadata

language:
  - ru
  - en
base_model: NousResearch/Llama-2-7b-hf
tags:
  - generated_from_trainer
  - bitnet
  - llama
  - rulm
  - darulm
datasets:
  - dichspace/darulm
library_name: transformers
model-index:
  - name: RuBit-Llama-56M2
    results: []

RuBit-Llama-63M

This model is a fine-tuned version of NousResearch/Llama-2-7b-hf on the darulm dataset. From darulm aphorisms, dramaturgy, history, humor, literature domains were sampled

Training on 2_125_871_104 tokens.

Inspired by abideen/Bitnet-Llama-70M

Model description

Sample inference code

from transformers import AutoModelForCausalLM, AutoTokenizer

# Load a pretrained BitNet model
model = "igorktech/RuBit-LLama-63M"
tokenizer = AutoTokenizer.from_pretrained(model)
model = AutoModelForCausalLM.from_pretrained(model)

def convert_to_bitnet(model, copy_weights):
    for name, module in model.named_modules():
        # Replace linear layers with BitNet
        if isinstance(module, LlamaSdpaAttention) or isinstance(module, LlamaMLP):
            for child_name, child_module in module.named_children():
                if isinstance(child_module, nn.Linear):
                    bitlinear = BitLinear(child_module.in_features, child_module.out_features, child_module.bias is not None).to(device="cuda:0")
                    if copy_weights:
                        bitlinear.weight = child_module.weight
                        if child_module.bias is not None:
                            bitlinear.bias = child_module.bias
                    setattr(module, child_name, bitlinear)
        # Remove redundant input_layernorms
        elif isinstance(module, LlamaDecoderLayer):
            for child_name, child_module in module.named_children():
                if isinstance(child_module, LlamaRMSNorm) and child_name == "input_layernorm":
                    setattr(module, child_name, nn.Identity().to(device="cuda:0"))
              

convert_to_bitnet(model, copy_weights=True)
model.to(device="cuda:0")

prompt = "Привет"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
generate_ids = model.generate(inputs.input_ids, max_length=100)
tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]

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: 0.0015
train_batch_size: 64
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 128
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 0.1
num_epochs: 2
mixed_precision_training: Native AMP

Training results

Framework versions

Transformers 4.40.0
Pytorch 2.2.1+cu121
Datasets 2.19.0
Tokenizers 0.19.1