Instructions to use tsavage68/Summary4500_M2_1000steps_1e7rate_05beta_CSFTDPO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use tsavage68/Summary4500_M2_1000steps_1e7rate_05beta_CSFTDPO with Transformers:

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

pipe = pipeline("text-generation", model="tsavage68/Summary4500_M2_1000steps_1e7rate_05beta_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_1000steps_1e7rate_05beta_CSFTDPO")
model = AutoModelForCausalLM.from_pretrained("tsavage68/Summary4500_M2_1000steps_1e7rate_05beta_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_1000steps_1e7rate_05beta_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_1000steps_1e7rate_05beta_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_1000steps_1e7rate_05beta_CSFTDPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/tsavage68/Summary4500_M2_1000steps_1e7rate_05beta_CSFTDPO

SGLang

How to use tsavage68/Summary4500_M2_1000steps_1e7rate_05beta_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_1000steps_1e7rate_05beta_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_1000steps_1e7rate_05beta_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_1000steps_1e7rate_05beta_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_1000steps_1e7rate_05beta_CSFTDPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use tsavage68/Summary4500_M2_1000steps_1e7rate_05beta_CSFTDPO with Docker Model Runner:
```
docker model run hf.co/tsavage68/Summary4500_M2_1000steps_1e7rate_05beta_CSFTDPO
```

Hyponatremia_M2_1000steps_1e7rate_05beta_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: 0.0014
Rewards/chosen: -1.1967
Rewards/rejected: -19.9702
Rewards/accuracies: 0.9980
Rewards/margins: 18.7736
Logps/rejected: -192.6703
Logps/chosen: -96.1331
Logits/rejected: -2.2591
Logits/chosen: -2.2130

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: 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: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Rewards/rejected	Rewards/accuracies	Rewards/margins	Logps/rejected	Logps/chosen	Logits/rejected	Logits/chosen
0.1654	0.0112	50	0.1737	-0.0508	-1.9689	0.9980	1.9181	-156.6675	-93.8414	-2.3390	-2.2918
0.0004	0.0224	100	0.0033	-0.7167	-11.2258	0.9980	10.5091	-175.1814	-95.1731	-2.2885	-2.2417
0.0	0.0336	150	0.0021	-1.0525	-14.0087	0.9980	12.9562	-180.7471	-95.8447	-2.2792	-2.2326
0.0	0.0448	200	0.0015	-0.9521	-16.6286	0.9980	15.6764	-185.9869	-95.6440	-2.2677	-2.2212
0.0	0.0559	250	0.0015	-0.9657	-17.2380	0.9980	16.2723	-187.2058	-95.6713	-2.2669	-2.2206
0.0	0.0671	300	0.0015	-0.9637	-17.2446	0.9980	16.2809	-187.2190	-95.6673	-2.2665	-2.2201
0.0	0.0783	350	0.0015	-1.1980	-18.6860	0.9980	17.4880	-190.1018	-96.1359	-2.2620	-2.2159
0.0001	0.0895	400	0.0014	-1.2301	-19.6059	0.9980	18.3757	-191.9415	-96.2000	-2.2577	-2.2117
0.0	0.1007	450	0.0015	-1.2380	-19.6415	0.9980	18.4035	-192.0128	-96.2158	-2.2573	-2.2113
0.0	0.1119	500	0.0014	-1.2365	-19.6568	0.9980	18.4203	-192.0434	-96.2128	-2.2581	-2.2121
0.0	0.1231	550	0.0014	-1.2308	-19.8868	0.9980	18.6559	-192.5033	-96.2015	-2.2587	-2.2127
0.0	0.1343	600	0.0014	-1.2131	-19.8634	0.9980	18.6504	-192.4567	-96.1659	-2.2581	-2.2121
0.0	0.1454	650	0.0014	-1.1869	-19.8805	0.9980	18.6936	-192.4907	-96.1136	-2.2606	-2.2145
0.0	0.1566	700	0.0014	-1.2139	-19.9693	0.9980	18.7554	-192.6684	-96.1675	-2.2588	-2.2127
0.0	0.1678	750	0.0014	-1.1965	-19.9802	0.9980	18.7837	-192.6902	-96.1328	-2.2595	-2.2134
0.0	0.1790	800	0.0014	-1.1843	-19.9036	0.9980	18.7193	-192.5370	-96.1084	-2.2606	-2.2145
0.0	0.1902	850	0.0014	-1.1914	-19.9692	0.9980	18.7778	-192.6682	-96.1225	-2.2591	-2.2130
0.0	0.2014	900	0.0014	-1.1979	-19.9798	0.9980	18.7819	-192.6894	-96.1356	-2.2589	-2.2128
0.0	0.2126	950	0.0014	-1.1962	-19.9695	0.9980	18.7733	-192.6688	-96.1321	-2.2591	-2.2130
0.0	0.2238	1000	0.0014	-1.1967	-19.9702	0.9980	18.7736	-192.6703	-96.1331	-2.2591	-2.2130

Framework versions

Transformers 4.42.4
Pytorch 2.0.0+cu117
Datasets 2.20.0
Tokenizers 0.19.1

Downloads last month: 6

Safetensors

Model size

7B params

Tensor type

F16

Model tree for tsavage68/Summary4500_M2_1000steps_1e7rate_05beta_CSFTDPO

Base model

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

Finetuned

tsavage68/Summary4500_M2_200steps_1e7rate_SFT

Finetuned

(14)

this model