Instructions to use tsavage68/Summary4500_M2_400steps_1e8rate_03beta_CSFTDPO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tsavage68/Summary4500_M2_400steps_1e8rate_03beta_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_03beta_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_03beta_CSFTDPO")
model = AutoModelForCausalLM.from_pretrained("tsavage68/Summary4500_M2_400steps_1e8rate_03beta_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_03beta_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_03beta_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_03beta_CSFTDPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tsavage68/Summary4500_M2_400steps_1e8rate_03beta_CSFTDPO

SGLang

How to use tsavage68/Summary4500_M2_400steps_1e8rate_03beta_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_03beta_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_03beta_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_03beta_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_03beta_CSFTDPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use tsavage68/Summary4500_M2_400steps_1e8rate_03beta_CSFTDPO with Docker Model Runner:
```
docker model run hf.co/tsavage68/Summary4500_M2_400steps_1e8rate_03beta_CSFTDPO
```

Hyponatremia_M2_400steps_1e8rate_03beta_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: 0.6713
Rewards/chosen: 0.0014
Rewards/rejected: -0.0515
Rewards/accuracies: 0.6260
Rewards/margins: 0.0529
Logps/rejected: -152.9013
Logps/chosen: -93.7350
Logits/rejected: -2.3519
Logits/chosen: -2.3045

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
0.6974	0.0112	50	0.6899	-0.0051	-0.0184	0.5020	0.0133	-152.7910	-93.7567	-2.3524	-2.3049
0.6814	0.0224	100	0.6951	-0.0040	-0.0071	0.5140	0.0032	-152.7536	-93.7530	-2.3533	-2.3059
0.6714	0.0336	150	0.6761	-0.0013	-0.0440	0.5880	0.0427	-152.8765	-93.7441	-2.3518	-2.3044
0.672	0.0448	200	0.6719	-0.0043	-0.0554	0.6160	0.0511	-152.9145	-93.7542	-2.3514	-2.3039
0.6919	0.0559	250	0.6657	-0.0015	-0.0659	0.6300	0.0644	-152.9496	-93.7449	-2.3512	-2.3038
0.6675	0.0671	300	0.6718	-0.0045	-0.0560	0.6120	0.0514	-152.9163	-93.7549	-2.3520	-2.3046
0.7033	0.0783	350	0.6714	0.0012	-0.0516	0.6260	0.0528	-152.9018	-93.7357	-2.3519	-2.3045
0.6112	0.0895	400	0.6713	0.0014	-0.0515	0.6260	0.0529	-152.9013	-93.7350	-2.3519	-2.3045

Framework versions

Transformers 4.42.4
Pytorch 2.0.0+cu117
Datasets 2.20.0
Tokenizers 0.19.1

Downloads last month: 5

Safetensors

Model size

7B params

Tensor type

F16

Model tree for tsavage68/Summary4500_M2_400steps_1e8rate_03beta_CSFTDPO

Base model

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

Finetuned

tsavage68/Summary4500_M2_200steps_1e7rate_SFT

Finetuned

(14)

this model