Instructions to use vinoku89/aimo3-exp-1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use vinoku89/aimo3-exp-1 with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-0.6B")
model = PeftModel.from_pretrained(base_model, "vinoku89/aimo3-exp-1")

Transformers

How to use vinoku89/aimo3-exp-1 with Transformers:

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

pipe = pipeline("text-generation", model="vinoku89/aimo3-exp-1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("vinoku89/aimo3-exp-1", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use vinoku89/aimo3-exp-1 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "vinoku89/aimo3-exp-1"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "vinoku89/aimo3-exp-1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/vinoku89/aimo3-exp-1

SGLang

How to use vinoku89/aimo3-exp-1 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 "vinoku89/aimo3-exp-1" \
    --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": "vinoku89/aimo3-exp-1",
		"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 "vinoku89/aimo3-exp-1" \
        --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": "vinoku89/aimo3-exp-1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use vinoku89/aimo3-exp-1 with Docker Model Runner:
```
docker model run hf.co/vinoku89/aimo3-exp-1
```

vinoku89 commited on Jan 12

Commit

bacfa81

verified ·

1 Parent(s): 09b3e6d

[sft] Qwen3-0.6B | e1 | 13.9k samples | lr=2e-05

Browse files

Metrics:
- train_loss: 2.0741
- val_loss: 2.0532

W&B: https://wandb.ai/run/79pv99nr

Files changed (3) hide show

adapter_config.json +5 -5
adapter_model.safetensors +1 -1
metrics.json +5 -5

adapter_config.json CHANGED Viewed

@@ -29,13 +29,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "up_proj",
-    "down_proj",
     "q_proj",
-    "k_proj",
     "o_proj",
-    "v_proj",
-    "gate_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

   "rank_pattern": {},
   "revision": null,
   "target_modules": [
+    "gate_proj",
+    "v_proj",
     "q_proj",
     "o_proj",
+    "down_proj",
+    "k_proj",
+    "up_proj"
   ],
   "target_parameters": null,
   "task_type": "CAUSAL_LM",

adapter_model.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f71d1008255502959aff19fed05c9f3a2605674476d1561d23721a240cb17022
 size 20237256

 version https://git-lfs.github.com/spec/v1
+oid sha256:cbd7f87d857fdcdd819cc88914741a22f095eb8ac3b9e5ea98a76d6ce13e3e3e
 size 20237256

metrics.json CHANGED Viewed

@@ -1,6 +1,6 @@
 {
-  "train_loss": 2.0751,
-  "val_loss": 2.0517611503601074,
   "accuracy": null,
   "epochs": 1,
   "max_steps": 10,
@@ -8,8 +8,8 @@
   "learning_rate": 2e-05,
   "model_name": "Qwen/Qwen3-0.6B",
   "method": "sft",
-  "run_number": "001",
   "experiment_name": "aimo3-exp-1",
-  "wandb_run_id": "3fkbe45r",
-  "timestamp": "2026-01-12T22:21:19.043295"
 }

 {
+  "train_loss": 2.0741,
+  "val_loss": 2.053159475326538,
   "accuracy": null,
   "epochs": 1,
   "max_steps": 10,
   "learning_rate": 2e-05,
   "model_name": "Qwen/Qwen3-0.6B",
   "method": "sft",
+  "run_number": "002",
   "experiment_name": "aimo3-exp-1",
+  "wandb_run_id": "79pv99nr",
+  "timestamp": "2026-01-12T22:26:23.825888"
 }