Instructions to use tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO with Transformers:

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

pipe = pipeline("text-generation", model="tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO")
model = AutoModelForCausalLM.from_pretrained("tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO", 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/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO"
# 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/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO

SGLang

How to use tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO 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/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO" \
    --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/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO",
		"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/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO" \
        --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/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO with Docker Model Runner:
```
docker model run hf.co/tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO
```

Hyponatremia_M2_400steps_1e8rate_01beta_CSFTDPO

This model is a fine-tuned version of tsavage68/Summary4500_M2_200steps_1e7rate_SFT on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.6838
Rewards/chosen: 0.0016
Rewards/rejected: -0.0180
Rewards/accuracies: 0.6400
Rewards/margins: 0.0196
Logps/rejected: -152.9099
Logps/chosen: -93.7234
Logits/rejected: -2.3527
Logits/chosen: -2.3052

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-08
train_batch_size: 1
eval_batch_size: 1
seed: 42
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
0.6837	0.0112	50	0.6927	-0.0020	-0.0036	0.4960	0.0017	-152.7662	-93.7594	-2.3523	-2.3049
0.6845	0.0224	100	0.6912	0.0015	-0.0032	0.5580	0.0047	-152.7614	-93.7247	-2.3530	-2.3055
0.693	0.0336	150	0.6873	-0.0010	-0.0136	0.6160	0.0126	-152.8659	-93.7498	-2.3518	-2.3044
0.6856	0.0448	200	0.6873	-0.0005	-0.0133	0.5920	0.0128	-152.8629	-93.7448	-2.3518	-2.3044
0.6791	0.0559	250	0.6863	-0.0009	-0.0155	0.5820	0.0146	-152.8851	-93.7492	-2.3522	-2.3047
0.6961	0.0671	300	0.6838	-0.0005	-0.0202	0.6320	0.0196	-152.9316	-93.7453	-2.3517	-2.3043
0.6984	0.0783	350	0.6840	0.0016	-0.0177	0.6380	0.0192	-152.9066	-93.7241	-2.3527	-2.3052
0.6724	0.0895	400	0.6838	0.0016	-0.0180	0.6400	0.0196	-152.9099	-93.7234	-2.3527	-2.3052

Framework versions

Transformers 4.42.4
Pytorch 2.0.0+cu117
Datasets 2.20.0
Tokenizers 0.19.1

Downloads last month: 4

Safetensors

Model size

7B params

Tensor type

F16

Model tree for tsavage68/Summary4500_M2_400steps_1e8rate_01beta_CSFTDPO

Base model

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

Finetuned

tsavage68/Summary4500_M2_200steps_1e7rate_SFT

Finetuned

(14)

this model