Instructions to use tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT with Transformers:

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

pipe = pipeline("text-generation", model="tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT")
model = AutoModelForCausalLM.from_pretrained("tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT

SGLang

How to use tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT 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/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT with Docker Model Runner:
```
docker model run hf.co/tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT
```

Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT

This model is a fine-tuned version of tsavage68/mistralit2_1000_STEPS_5e7_SFT on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.5720
Rewards/chosen: 0.2834
Rewards/rejected: -0.1194
Rewards/accuracies: 0.6484
Rewards/margins: 0.4028
Logps/rejected: -27.7509
Logps/chosen: -20.8379
Logits/rejected: -2.1751
Logits/chosen: -2.1746

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: 300

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.6891	0.0977	50	0.6864	0.0309	0.0171	0.6220	0.0138	-26.3858	-23.3627	-2.3067	-2.3063
0.6478	0.1953	100	0.6468	0.0940	-0.0114	0.6418	0.1054	-26.6706	-22.7312	-2.2655	-2.2651
0.5684	0.2930	150	0.5979	0.1517	-0.1062	0.6593	0.2579	-27.6188	-22.1551	-2.2170	-2.2166
0.4682	0.3906	200	0.5760	0.2970	-0.0758	0.6462	0.3728	-27.3147	-20.7013	-2.1834	-2.1830
0.5035	0.4883	250	0.5718	0.2838	-0.1177	0.6484	0.4015	-27.7339	-20.8337	-2.1751	-2.1747
0.4736	0.5859	300	0.5720	0.2834	-0.1194	0.6484	0.4028	-27.7509	-20.8379	-2.1751	-2.1746

Framework versions

Transformers 4.40.1
Pytorch 2.0.0+cu117
Datasets 2.19.0
Tokenizers 0.19.1

Downloads last month: 8

Safetensors

Model size

7B params

Tensor type

F16

Model tree for tsavage68/Mistral2_300_STEPS_01beta_1e7rate_CDPOSFT

Base model

mistralai/Mistral-7B-Instruct-v0.2

Finetuned

tsavage68/mistralit2_1000_STEPS_5e7_SFT

Finetuned

(19)

this model