Instructions to use allura-org/Koto-Small-7B-PT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use allura-org/Koto-Small-7B-PT with Transformers:

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

pipe = pipeline("text-generation", model="allura-org/Koto-Small-7B-PT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("allura-org/Koto-Small-7B-PT")
model = AutoModelForCausalLM.from_pretrained("allura-org/Koto-Small-7B-PT")
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

vLLM

How to use allura-org/Koto-Small-7B-PT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "allura-org/Koto-Small-7B-PT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "allura-org/Koto-Small-7B-PT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/allura-org/Koto-Small-7B-PT

SGLang

How to use allura-org/Koto-Small-7B-PT 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 "allura-org/Koto-Small-7B-PT" \
    --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": "allura-org/Koto-Small-7B-PT",
		"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 "allura-org/Koto-Small-7B-PT" \
        --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": "allura-org/Koto-Small-7B-PT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use allura-org/Koto-Small-7B-PT with Docker Model Runner:
```
docker model run hf.co/allura-org/Koto-Small-7B-PT
```

Fizzarolli commited on Aug 31, 2025

Commit

2e743e5

verified ·

1 Parent(s): 6c1bfe3

Update README.md

Browse files

Files changed (1) hide show

README.md +2 -0

README.md CHANGED Viewed

@@ -52,6 +52,8 @@ please join [our discord](https://discord.gg/PPBMhF2vgC) or [our matrix](https:/
 This model was trained over the course of ~18 hours on an A100 node. We used 8-bit AdamW and the Cosine LR scheduler, as well as both gradient clipping and weight decay for regularization.
 Before training, we [converted the original model to the Qwen 2 architecture](https://huggingface.co/allura-forge/MiMo-7B-Base-Qwenified) by removing the MTP weights and custom modelling code, and slightly modifying the `config.json`. This allowed us to use CCE and Liger which let the train go much faster than it would have otherwise.
 ### [WandB](https://wandb.ai/new-eden/Koto-Small/runs/zk8t6oq6/workspace)
 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/634262af8d8089ebaefd410e/Fc-Dvakg3lSwk2co7jHIM.png)

 This model was trained over the course of ~18 hours on an A100 node. We used 8-bit AdamW and the Cosine LR scheduler, as well as both gradient clipping and weight decay for regularization.
 Before training, we [converted the original model to the Qwen 2 architecture](https://huggingface.co/allura-forge/MiMo-7B-Base-Qwenified) by removing the MTP weights and custom modelling code, and slightly modifying the `config.json`. This allowed us to use CCE and Liger which let the train go much faster than it would have otherwise.
+We decided to keep the final model in the converted Qwen 2 format, as it is more supported by community software such as EXL2, EXL3, Aphrodite, etc, as well as the original architecture's MTP weights likely being much less effective after finetuning without them.
 ### [WandB](https://wandb.ai/new-eden/Koto-Small/runs/zk8t6oq6/workspace)
 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/634262af8d8089ebaefd410e/Fc-Dvakg3lSwk2co7jHIM.png)