Instructions to use YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03 with Transformers:

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

pipe = pipeline("text-generation", model="YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03")
model = AutoModelForCausalLM.from_pretrained("YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03", 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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03

SGLang

How to use YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03 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 "YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03" \
    --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": "YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03",
		"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 "YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03" \
        --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": "YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03 with Docker Model Runner:
```
docker model run hf.co/YeongminKim/zephyr-7b-dpo-full-alpha_0.5_batch64_0.03
```

YeongminKim commited on Apr 21, 2025

Commit

4bf1a1a

verified ·

1 Parent(s): 00937f2

End of training

Browse files

Files changed (4) hide show

README.md +15 -9
all_results.json +13 -0
config.json +1 -1
eval_results.json +16 -0

README.md CHANGED Viewed

@@ -3,9 +3,15 @@ library_name: transformers
 license: apache-2.0
 base_model: alignment-handbook/zephyr-7b-sft-full
 tags:
 - trl
 - dpo
 - generated_from_trainer
 model-index:
 - name: zephyr-7b-dpo-full-alpha_0.5_batch64_0.03
   results: []
@@ -16,17 +22,17 @@ should probably proofread and complete it, then remove this comment. -->
 # zephyr-7b-dpo-full-alpha_0.5_batch64_0.03
-This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the None dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.7788
-- Rewards/chosen: -1.3317
-- Rewards/rejected: -2.3491
 - Rewards/accuracies: 0.7857
-- Rewards/margins: 1.0174
-- Logps/rejected: -495.1149
-- Logps/chosen: -415.1518
-- Logits/rejected: 0.3728
-- Logits/chosen: -0.7858
 ## Model description

 license: apache-2.0
 base_model: alignment-handbook/zephyr-7b-sft-full
 tags:
+- alignment-handbook
 - trl
 - dpo
 - generated_from_trainer
+- trl
+- dpo
+- generated_from_trainer
+datasets:
+- HuggingFaceH4/ultrafeedback_binarized
 model-index:
 - name: zephyr-7b-dpo-full-alpha_0.5_batch64_0.03
   results: []
 # zephyr-7b-dpo-full-alpha_0.5_batch64_0.03
+This model is a fine-tuned version of [alignment-handbook/zephyr-7b-sft-full](https://huggingface.co/alignment-handbook/zephyr-7b-sft-full) on the HuggingFaceH4/ultrafeedback_binarized dataset.
 It achieves the following results on the evaluation set:
+- Loss: 0.7789
+- Rewards/chosen: -1.3273
+- Rewards/rejected: -2.3422
 - Rewards/accuracies: 0.7857
+- Rewards/margins: 1.0149
+- Logps/rejected: -494.4228
+- Logps/chosen: -414.7076
+- Logits/rejected: 0.3577
+- Logits/chosen: -0.7982
 ## Model description

all_results.json CHANGED Viewed

@@ -1,5 +1,18 @@
 {
     "epoch": 0.9994767137624281,
     "total_flos": 0.0,
     "train_loss": 0.8168348412239115,
     "train_runtime": 34624.4791,

 {
     "epoch": 0.9994767137624281,
+    "eval_logits/chosen": -0.7981740236282349,
+    "eval_logits/rejected": 0.35771894454956055,
+    "eval_logps/chosen": -414.70758056640625,
+    "eval_logps/rejected": -494.42279052734375,
+    "eval_loss": 0.7788758277893066,
+    "eval_rewards/accuracies": 0.7857142686843872,
+    "eval_rewards/chosen": -1.3273053169250488,
+    "eval_rewards/margins": 1.0148944854736328,
+    "eval_rewards/rejected": -2.3421998023986816,
+    "eval_runtime": 408.2742,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 4.899,
+    "eval_steps_per_second": 0.154,
     "total_flos": 0.0,
     "train_loss": 0.8168348412239115,
     "train_runtime": 34624.4791,

config.json CHANGED Viewed

@@ -22,6 +22,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.2",
-  "use_cache": false,
   "vocab_size": 32000
 }

   "tie_word_embeddings": false,
   "torch_dtype": "bfloat16",
   "transformers_version": "4.44.2",
+  "use_cache": true,
   "vocab_size": 32000
 }

eval_results.json ADDED Viewed

	@@ -0,0 +1,16 @@

+{
+    "epoch": 0.9994767137624281,
+    "eval_logits/chosen": -0.7981740236282349,
+    "eval_logits/rejected": 0.35771894454956055,
+    "eval_logps/chosen": -414.70758056640625,
+    "eval_logps/rejected": -494.42279052734375,
+    "eval_loss": 0.7788758277893066,
+    "eval_rewards/accuracies": 0.7857142686843872,
+    "eval_rewards/chosen": -1.3273053169250488,
+    "eval_rewards/margins": 1.0148944854736328,
+    "eval_rewards/rejected": -2.3421998023986816,
+    "eval_runtime": 408.2742,
+    "eval_samples": 2000,
+    "eval_samples_per_second": 4.899,
+    "eval_steps_per_second": 0.154
+}