Instructions to use melino2000/tiny-llama-merge-evo with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use melino2000/tiny-llama-merge-evo with Transformers:

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

pipe = pipeline("text-generation", model="melino2000/tiny-llama-merge-evo")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("melino2000/tiny-llama-merge-evo")
model = AutoModelForCausalLM.from_pretrained("melino2000/tiny-llama-merge-evo")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use melino2000/tiny-llama-merge-evo with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "melino2000/tiny-llama-merge-evo"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "melino2000/tiny-llama-merge-evo",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/melino2000/tiny-llama-merge-evo

SGLang

How to use melino2000/tiny-llama-merge-evo 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 "melino2000/tiny-llama-merge-evo" \
    --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": "melino2000/tiny-llama-merge-evo",
		"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 "melino2000/tiny-llama-merge-evo" \
        --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": "melino2000/tiny-llama-merge-evo",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use melino2000/tiny-llama-merge-evo with Docker Model Runner:
```
docker model run hf.co/melino2000/tiny-llama-merge-evo
```

final_merge

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the task arithmetic merge method using /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064 as a base.

Models Merged

The following models were included in the merge:

/kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950

Configuration

The following YAML configuration was used to produce this model:

base_model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
dtype: bfloat16
merge_method: task_arithmetic
parameters:
  int8_mask: 1.0
  normalize: 0.0
slices:
- sources:
  - layer_range: [0, 2]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 0.903670769683462
  - layer_range: [0, 2]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
- sources:
  - layer_range: [2, 4]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 0.8677123591929141
  - layer_range: [2, 4]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
- sources:
  - layer_range: [4, 6]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 1.0080967885131624
  - layer_range: [4, 6]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
- sources:
  - layer_range: [6, 8]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 1.288794492088366
  - layer_range: [6, 8]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
- sources:
  - layer_range: [8, 10]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 1.0927250789898328
  - layer_range: [8, 10]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
- sources:
  - layer_range: [10, 12]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 1.002818025226096
  - layer_range: [10, 12]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
- sources:
  - layer_range: [12, 14]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 1.0346267702747531
  - layer_range: [12, 14]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
- sources:
  - layer_range: [14, 16]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 1.060865068400883
  - layer_range: [14, 16]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
- sources:
  - layer_range: [16, 18]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 1.0141257624580193
  - layer_range: [16, 18]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
- sources:
  - layer_range: [18, 20]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 1.2727977176081706
  - layer_range: [18, 20]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064
- sources:
  - layer_range: [20, 22]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_chinese_1905817950
    parameters:
      weight: 1.2137521068579595
  - layer_range: [20, 22]
    model: /kaggle/working/evol_merge_storage/input_models/TinyLlama_v1.1_684560064

Downloads last month: 1

Safetensors

Model size

1B params

Tensor type

BF16

Paper for melino2000/tiny-llama-merge-evo

Editing Models with Task Arithmetic

Paper • 2212.04089 • Published Dec 8, 2022 • 9