Instructions to use tsavage68/400STEPS_1e7rate_01beta_Mistral with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tsavage68/400STEPS_1e7rate_01beta_Mistral with Transformers:

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

pipe = pipeline("text-generation", model="tsavage68/400STEPS_1e7rate_01beta_Mistral")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tsavage68/400STEPS_1e7rate_01beta_Mistral")
model = AutoModelForCausalLM.from_pretrained("tsavage68/400STEPS_1e7rate_01beta_Mistral", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tsavage68/400STEPS_1e7rate_01beta_Mistral with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tsavage68/400STEPS_1e7rate_01beta_Mistral"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tsavage68/400STEPS_1e7rate_01beta_Mistral",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/tsavage68/400STEPS_1e7rate_01beta_Mistral

SGLang

How to use tsavage68/400STEPS_1e7rate_01beta_Mistral 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 "tsavage68/400STEPS_1e7rate_01beta_Mistral" \
    --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": "tsavage68/400STEPS_1e7rate_01beta_Mistral",
		"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 "tsavage68/400STEPS_1e7rate_01beta_Mistral" \
        --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": "tsavage68/400STEPS_1e7rate_01beta_Mistral",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use tsavage68/400STEPS_1e7rate_01beta_Mistral with Docker Model Runner:
```
docker model run hf.co/tsavage68/400STEPS_1e7rate_01beta_Mistral
```

400STEPS_1e7rate_01beta_T5

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.6483
Rewards/chosen: -0.0026
Rewards/rejected: -0.1019
Rewards/accuracies: 0.6593
Rewards/margins: 0.0994
Logps/rejected: -15.7387
Logps/chosen: -12.9908
Logits/rejected: -3.1652
Logits/chosen: -3.1650

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-07
train_batch_size: 4
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 100
training_steps: 400

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.6916	0.1	50	0.6908	0.0048	0.0002	0.5670	0.0047	-14.7176	-12.9168	-3.1591	-3.1588
0.6821	0.2	100	0.6764	0.0187	-0.0159	0.6681	0.0346	-14.8782	-12.7778	-3.1625	-3.1622
0.6647	0.29	150	0.6629	0.0225	-0.0422	0.6659	0.0648	-15.1414	-12.7399	-3.1625	-3.1623
0.6536	0.39	200	0.6552	0.0148	-0.0679	0.6505	0.0827	-15.3987	-12.8175	-3.1657	-3.1654
0.6354	0.49	250	0.6509	0.0022	-0.0909	0.6593	0.0931	-15.6282	-12.9431	-3.1646	-3.1643
0.6468	0.59	300	0.6484	-0.0022	-0.1013	0.6527	0.0991	-15.7319	-12.9869	-3.1653	-3.1650
0.6549	0.68	350	0.6481	-0.0021	-0.1019	0.6571	0.0998	-15.7386	-12.9865	-3.1652	-3.1650
0.6684	0.78	400	0.6483	-0.0026	-0.1019	0.6593	0.0994	-15.7387	-12.9908	-3.1652	-3.1650

Framework versions

Transformers 4.37.2
Pytorch 2.0.0+cu117
Datasets 2.17.0
Tokenizers 0.15.1

Downloads last month: 5

Safetensors

Model size

7B params

Tensor type

F16

Model tree for tsavage68/400STEPS_1e7rate_01beta_Mistral

Base model

mistralai/Mistral-7B-v0.1

Finetuned

(945)

this model