Instructions to use unsloth/codellama-34b-bnb-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use unsloth/codellama-34b-bnb-4bit with Transformers:

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

pipe = pipeline("text-generation", model="unsloth/codellama-34b-bnb-4bit")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("unsloth/codellama-34b-bnb-4bit")
model = AutoModelForCausalLM.from_pretrained("unsloth/codellama-34b-bnb-4bit")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use unsloth/codellama-34b-bnb-4bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "unsloth/codellama-34b-bnb-4bit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "unsloth/codellama-34b-bnb-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/unsloth/codellama-34b-bnb-4bit

SGLang

How to use unsloth/codellama-34b-bnb-4bit 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 "unsloth/codellama-34b-bnb-4bit" \
    --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": "unsloth/codellama-34b-bnb-4bit",
		"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 "unsloth/codellama-34b-bnb-4bit" \
        --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": "unsloth/codellama-34b-bnb-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Unsloth Studio

How to use unsloth/codellama-34b-bnb-4bit with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for unsloth/codellama-34b-bnb-4bit to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for unsloth/codellama-34b-bnb-4bit to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for unsloth/codellama-34b-bnb-4bit to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="unsloth/codellama-34b-bnb-4bit",
    max_seq_length=2048,
)

Docker Model Runner
How to use unsloth/codellama-34b-bnb-4bit with Docker Model Runner:
```
docker model run hf.co/unsloth/codellama-34b-bnb-4bit
```

danielhanchen commited on Aug 7, 2024

Commit

0a08e04

verified ·

1 Parent(s): 6b5fd95

Upload config

Browse files

Files changed (2) hide show

README.md +1 -2
config.json +8 -3

README.md CHANGED Viewed

@@ -1,14 +1,13 @@
 ---
 language:
 - en
-license: apache-2.0
 library_name: transformers
 tags:
 - unsloth
 - transformers
 - codellama
 - codellama-34b
 ---
 # Finetune Mistral, Gemma, Llama 2-5x faster with 70% less memory via Unsloth!

 ---
 language:
 - en
 library_name: transformers
+license: apache-2.0
 tags:
 - unsloth
 - transformers
 - codellama
 - codellama-34b
 ---
 # Finetune Mistral, Gemma, Llama 2-5x faster with 70% less memory via Unsloth!

config.json CHANGED Viewed

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "codellama/CodeLlama-34b-hf",
   "architectures": [
     "LlamaForCausalLM"
   ],
@@ -12,14 +12,18 @@
   "initializer_range": 0.02,
   "intermediate_size": 22016,
   "max_position_embeddings": 16384,
   "model_type": "llama",
   "num_attention_heads": 64,
   "num_hidden_layers": 48,
   "num_key_value_heads": 8,
-  "pad_token_id": 2,
   "pretraining_tp": 1,
   "quantization_config": {
     "bnb_4bit_compute_dtype": "bfloat16",
     "bnb_4bit_quant_type": "nf4",
     "bnb_4bit_use_double_quant": true,
     "llm_int8_enable_fp32_cpu_offload": false,
@@ -35,7 +39,8 @@
   "rope_theta": 1000000,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
-  "transformers_version": "4.37.0.dev0",
   "use_cache": true,
   "vocab_size": 32000
 }

 {
+  "_name_or_path": "unsloth/codellama-34b",
   "architectures": [
     "LlamaForCausalLM"
   ],
   "initializer_range": 0.02,
   "intermediate_size": 22016,
   "max_position_embeddings": 16384,
+  "mlp_bias": false,
   "model_type": "llama",
   "num_attention_heads": 64,
   "num_hidden_layers": 48,
   "num_key_value_heads": 8,
+  "pad_token_id": 0,
   "pretraining_tp": 1,
   "quantization_config": {
+    "_load_in_4bit": true,
+    "_load_in_8bit": false,
     "bnb_4bit_compute_dtype": "bfloat16",
+    "bnb_4bit_quant_storage": "uint8",
     "bnb_4bit_quant_type": "nf4",
     "bnb_4bit_use_double_quant": true,
     "llm_int8_enable_fp32_cpu_offload": false,
   "rope_theta": 1000000,
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
+  "transformers_version": "4.44.0",
+  "unsloth_version": "2024.8",
   "use_cache": true,
   "vocab_size": 32000
 }