Instructions to use Chat-Error/Kimiko-Mistral-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Chat-Error/Kimiko-Mistral-7B with Transformers:

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

pipe = pipeline("text-generation", model="Chat-Error/Kimiko-Mistral-7B")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Chat-Error/Kimiko-Mistral-7B")
model = AutoModelForCausalLM.from_pretrained("Chat-Error/Kimiko-Mistral-7B")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Chat-Error/Kimiko-Mistral-7B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Chat-Error/Kimiko-Mistral-7B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Chat-Error/Kimiko-Mistral-7B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Chat-Error/Kimiko-Mistral-7B

SGLang

How to use Chat-Error/Kimiko-Mistral-7B 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 "Chat-Error/Kimiko-Mistral-7B" \
    --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": "Chat-Error/Kimiko-Mistral-7B",
		"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 "Chat-Error/Kimiko-Mistral-7B" \
        --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": "Chat-Error/Kimiko-Mistral-7B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Chat-Error/Kimiko-Mistral-7B with Docker Model Runner:
```
docker model run hf.co/Chat-Error/Kimiko-Mistral-7B
```

Kimiko-Mistral-7B

(I am going to retrain this, this model is a failure) This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the Kimiko dataset. It achieves the following results on the evaluation set:

Loss: 2.1173

Model description

Same dataset as Kimiko-v2 but on new model. THIS IS NOT TRAIN ON V3 DATASET

Intended uses & limitations

As a finetuning experiment on new 7B model. You can use this for roleplay or as an assistant

Prompt Template Structure

This is a chat between ASSISTANT and USER
USER: What is 4x8?
ASSISTANT:

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.00005
train_batch_size: 4
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 16
total_train_batch_size: 64
optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-05
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 10
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
1.5675	0.47	25	2.1323
1.4721	0.95	50	2.1209
1.472	1.42	75	2.1177
1.5445	1.9	100	2.1173

Framework versions

Transformers 4.34.0.dev0
Pytorch 2.0.1+cu118
Datasets 2.14.5
Tokenizers 0.14.0

Downloads last month: 10

Model tree for Chat-Error/Kimiko-Mistral-7B

Base model

mistralai/Mistral-7B-v0.1

Quantized

(184)

this model

Finetunes

1 model

Quantizations

3 models