Instructions to use Kortix/FastApply-1.5B-v1.0 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Kortix/FastApply-1.5B-v1.0 with Transformers:

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

pipe = pipeline("text-generation", model="Kortix/FastApply-1.5B-v1.0")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Kortix/FastApply-1.5B-v1.0")
model = AutoModelForCausalLM.from_pretrained("Kortix/FastApply-1.5B-v1.0")
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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Kortix/FastApply-1.5B-v1.0 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Kortix/FastApply-1.5B-v1.0"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Kortix/FastApply-1.5B-v1.0",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Kortix/FastApply-1.5B-v1.0

SGLang

How to use Kortix/FastApply-1.5B-v1.0 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 "Kortix/FastApply-1.5B-v1.0" \
    --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": "Kortix/FastApply-1.5B-v1.0",
		"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 "Kortix/FastApply-1.5B-v1.0" \
        --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": "Kortix/FastApply-1.5B-v1.0",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Studio new

How to use Kortix/FastApply-1.5B-v1.0 with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for Kortix/FastApply-1.5B-v1.0 to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for Kortix/FastApply-1.5B-v1.0 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Kortix/FastApply-1.5B-v1.0 to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="Kortix/FastApply-1.5B-v1.0",
    max_seq_length=2048,
)

Docker Model Runner
How to use Kortix/FastApply-1.5B-v1.0 with Docker Model Runner:
```
docker model run hf.co/Kortix/FastApply-1.5B-v1.0
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

FastApply-1.5B-v1.0

Github: kortix-ai/fast-apply
Dataset: Kortix/FastApply-dataset-v1.0
Try it now on 👉 Google Colab

Model Details

Basic Information

Developed by: Kortix
License: apache-2.0
Finetuned from model: unsloth/Qwen2.5-Coder-1.5B-Instruct-bnb-4bit

Model Description

FastApply-1.5B-v1.0 is a 1.5B model designed for instant code application, producing full file edits to power SoftGen AI.
It is part of the Fast Apply pipeline for data generation and fine-tuning Qwen2.5 Coder models.

The model achieves high throughput when deployed on fast providers like Fireworks while maintaining high edit accuracy, with a speed of approximately 340 tokens/second.

Intended Use

FastApply-1.5B-v1.0 is intended for use in AI-powered code editors and tools that require fast, accurate code modifications. It is particularly well-suited for:

Instant code application tasks
Full file edits
Integration with AI-powered code editors like Aider and PearAI
Local tools to reduce the cost of frontier model output

Inference template

FastApply-1.5B-v1.0 is based on the Qwen2.5 Coder architecture and is fine-tuned for code editing tasks. It uses a specific prompt structure for inference:

<|im_start|>system
You are a coding assistant that helps merge code updates, ensuring every modification is fully integrated.<|im_end|>
<|im_start|>user
Merge all changes from the <update> snippet into the <code> below.
- Preserve the code's structure, order, comments, and indentation exactly.
- Output only the updated code, enclosed within <updated-code> and </updated-code> tags.
- Do not include any additional text, explanations, placeholders, ellipses, or code fences.

<code>{original_code}</code>

<update>{update_snippet}</update>

Provide the complete updated code.<|im_end|>
<|im_start|>assistant

The model's output is structured as:

<updated-code>[Full-complete updated file]</updated-code>

Additional Information

For more details on the Fast Apply pipeline, data generation process, and deployment instructions, please refer to the GitHub repository.

How to Use

To use the model, you can load it using the Hugging Face Transformers library:

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("Kortix/FastApply-1.5B-v1.0", device_map="auto")
tokenizer = AutoTokenizer.from_pretrained("Kortix/FastApply-1.5B-v1.0")

# Prepare your input following the prompt structure mentioned above
input_text = """<|im_start|>system
You are a coding assistant that helps merge code updates, ensuring every modification is fully integrated.<|im_end|>
<|im_start|>user
Merge all changes from the <update> snippet into the <code> below.
- Preserve the code's structure, order, comments, and indentation exactly.
- Output only the updated code, enclosed within <updated-code> and </updated-code> tags.
- Do not include any additional text, explanations, placeholders, ellipses, or code fences.

<code>{original_code}</code>

<update>{update_snippet}</update>

Provide the complete updated code.<|im_end|>
<|im_start|>assistant
"""

input_text = input_text.format(
    original_code=original_code,
    update_snippet=update_snippet,
).strip() 

# Generate the response
input_ids = tokenizer.encode(input_text, return_tensors="pt")
output = model.generate(input_ids, max_length=8192,)

response = tokenizer.decode(output[0][len(input_ids[0]):])
print(response)

# Extract the updated code from the response
updated_code = response.split("<updated-code>")[1].split("</updated-code>")[0]

Evaluation:

Downloads last month: 126

Safetensors

Model size

2B params

Tensor type

BF16

Model tree for Kortix/FastApply-1.5B-v1.0

Finetunes

2 models

Quantizations

15 models