# /// script # requires-python = ">=3.10,<3.14" # dependencies = [ # "bitsandbytes>=0.43.0", # "marimo", # "tokenizers>=0.22.0,<=0.23.0", # "torch>=2.8.0", # "torchao>=0.16.0", # "torchvision", # "transformers==4.56.2", # "triton>=3.2.0", # "triton_kernels @ git+https://github.com/triton-lang/triton.git@0add68262ab0a2e33b84524346cb27cbb2787356#subdirectory=python/triton_kernels", # "trl==0.22.2", # "unsloth @ git+https://github.com/unslothai/unsloth", # "unsloth_zoo @ git+https://github.com/unslothai/unsloth-zoo", # "uv", # ] # # [tool.uv] # no-build-package = [ # "bitsandbytes", # "triton", # "vllm", # "xformers", # ] # /// import marimo __generated_with = "0.23.8" app = marimo.App() @app.cell def _(): import marimo as mo return (mo,) @app.cell(hide_code=True) def _(mo): mo.md(r""" To run this notebook, hit the **▶ Run all** button in the bottom-right corner - or use `Ctrl/Cmd + Shift + R`.

Join Discord if you need help + ⭐ Star us on Github ⭐

To install Unsloth on your local device, follow [our guide](https://unsloth.ai/docs/get-started/install). This notebook is licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme). You will learn how to do [data prep](#Data), how to [train](#Train), how to [run the model](#Inference), & how to save it """) return @app.cell(hide_code=True) def _(mo): mo.md(r""" ### News """) return @app.cell(hide_code=True) def _(mo): mo.md(r""" Introducing **Unsloth Studio** - a new open source, no-code web UI to train and run LLMs. [Blog](https://unsloth.ai/docs/new/studio) • [Notebook](https://github.com/unslothai/unsloth/blob/main/studio/Unsloth_Studio_Colab.ipynb)

_{Train models — no code needed}

_{Run GGUF models on Mac, Windows & Linux}

Train MoEs - DeepSeek, GLM, Qwen and gpt-oss 12x faster with 35% less VRAM. [Blog](https://unsloth.ai/docs/new/faster-moe) Ultra Long-Context Reinforcement Learning is here with 7x more context windows! [Blog](https://unsloth.ai/docs/new/grpo-long-context) New in Reinforcement Learning: [FP8 RL](https://unsloth.ai/docs/new/fp8-reinforcement-learning) • [Vision RL](https://unsloth.ai/docs/new/vision-reinforcement-learning-vlm-rl) • [Standby](https://unsloth.ai/docs/basics/memory-efficient-rl) • [gpt-oss RL](https://unsloth.ai/docs/new/gpt-oss-reinforcement-learning) Visit our docs for all our [model uploads](https://unsloth.ai/docs/get-started/unsloth-model-catalog) and [notebooks](https://unsloth.ai/docs/get-started/unsloth-notebooks). """) return @app.cell(hide_code=True) def _(mo): mo.md(r""" ### Unsloth """) return @app.cell(hide_code=True) def _(mo): mo.md(r""" We're about to demonstrate the power of the new OpenAI GPT-OSS 20B model through a finetuning example. To use our `MXFP4` inference example, use this [notebook](https://github.com/unslothai/notebooks/blob/main/nb/GPT_OSS_MXFP4_(20B)-Inference.ipynb) instead. """) return @app.cell def _(): from unsloth import FastLanguageModel import torch max_seq_length = 1024 # Choose any for long context! dtype = None # None for auto detection # 4bit pre quantized models we support for 4x faster downloading + no OOMs. fourbit_models = [ "unsloth/gpt-oss-20b-unsloth-bnb-4bit", # 20B model using bitsandbytes 4bit quantization "unsloth/gpt-oss-120b-unsloth-bnb-4bit", "unsloth/gpt-oss-20b", # 20B model using MXFP4 format "unsloth/gpt-oss-120b", ] # More models at https://huggingface.co/unsloth model, tokenizer = FastLanguageModel.from_pretrained( model_name="unsloth/gpt-oss-20b", # YOUR MODEL YOU USED FOR TRAINING dtype=dtype, # None for auto detection max_seq_length=max_seq_length, # Choose any for long context! load_in_4bit=True, # 4 bit quantization to reduce memory full_finetuning=False, # [NEW!] We have full finetuning now! # token = "YOUR_HF_TOKEN", # HF Token for gated models ) return FastLanguageModel, model, tokenizer, torch @app.cell(hide_code=True) def _(mo): mo.md(r""" We now add LoRA adapters for parameter efficient finetuning - this allows us to only efficiently train 1% of all parameters. """) return @app.cell def _(FastLanguageModel, model): model_1 = FastLanguageModel.get_peft_model( model, r=8, # Choose any number > 0 ! Suggested 8, 16, 32, 64, 128 target_modules=[ "q_proj", "k_proj", "v_proj", "o_proj", "gate_proj", "up_proj", "down_proj", ], lora_alpha=16, lora_dropout=0, # Supports any, but = 0 is optimized bias="none", # Supports any, but = "none" is optimized use_gradient_checkpointing="unsloth", # True or "unsloth" for very long context random_state=3407, use_rslora=False, # We support rank stabilized LoRA loftq_config=None, # And LoftQ ) return (model_1,) @app.cell(hide_code=True) def _(mo): mo.md(r""" ### Reasoning Effort The `gpt-oss` models from OpenAI include a feature that allows users to adjust the model's "reasoning effort." This gives you control over the trade-off between the model's performance and its response speed (latency) which by the amount of token the model will use to think. ---- The `gpt-oss` models offer three distinct levels of reasoning effort you can choose from: * **Low**: Optimized for tasks that need very fast responses and don't require complex, multi-step reasoning. * **Medium**: A balance between performance and speed. * **High**: Provides the strongest reasoning performance for tasks that require it, though this results in higher latency. """) return @app.cell def _(model_1, tokenizer): from transformers import TextStreamer messages = [{"role": "user", "content": "Solve x^5 + 3x^4 - 10 = 3."}] inputs = tokenizer.apply_chat_template( messages, add_generation_prompt=True, return_tensors="pt", return_dict=True, reasoning_effort="low", # **NEW!** Set reasoning effort to low, medium or high ).to("cuda") _ = model_1.generate( **inputs, max_new_tokens=64, streamer=TextStreamer(tokenizer) ) # **NEW!** Set reasoning effort to low, medium or high return (TextStreamer,) @app.cell(hide_code=True) def _(mo): mo.md(r""" Changing the `reasoning_effort` to `medium` will make the model think longer. We have to increase the `max_new_tokens` to occupy the amount of the generated tokens but it will give better and more correct answer """) return @app.cell def _(TextStreamer, model_1, tokenizer): messages_1 = [{"role": "user", "content": "Solve x^5 + 3x^4 - 10 = 3."}] inputs_1 = tokenizer.apply_chat_template( messages_1, add_generation_prompt=True, return_tensors="pt", return_dict=True, reasoning_effort="medium", # **NEW!** Set reasoning effort to low, medium or high ).to("cuda") _ = model_1.generate( **inputs_1, max_new_tokens=64, streamer=TextStreamer(tokenizer) ) # **NEW!** Set reasoning effort to low, medium or high return @app.cell(hide_code=True) def _(mo): mo.md(r""" Lastly we will test it using `reasoning_effort` to `high` """) return @app.cell def _(TextStreamer, model_1, tokenizer): messages_2 = [{"role": "user", "content": "Solve x^5 + 3x^4 - 10 = 3."}] inputs_2 = tokenizer.apply_chat_template( messages_2, add_generation_prompt=True, return_tensors="pt", return_dict=True, reasoning_effort="high", # **NEW!** Set reasoning effort to low, medium or high ).to("cuda") _ = model_1.generate( **inputs_2, max_new_tokens=64, streamer=TextStreamer(tokenizer) ) # **NEW!** Set reasoning effort to low, medium or high return @app.cell(hide_code=True) def _(mo): mo.md(r""" ### Data Prep """) return @app.cell(hide_code=True) def _(mo): mo.md(r""" The `HuggingFaceH4/Multilingual-Thinking` dataset will be utilized as our example. This dataset, available on Hugging Face, contains reasoning chain-of-thought examples derived from user questions that have been translated from English into four other languages. It is also the same dataset referenced in OpenAI's [cookbook](https://cookbook.openai.com/articles/gpt-oss/fine-tune-transfomers) for fine-tuning. The purpose of using this dataset is to enable the model to learn and develop reasoning capabilities in these four distinct languages. """) return @app.cell def _(tokenizer): def formatting_prompts_func(examples): convos = examples["messages"] texts = [ tokenizer.apply_chat_template( convo, tokenize=False, add_generation_prompt=False ) for convo in convos ] return { "text": texts, } from datasets import load_dataset dataset = load_dataset("HuggingFaceH4/Multilingual-Thinking", split="train") dataset return dataset, formatting_prompts_func @app.cell(hide_code=True) def _(mo): mo.md(r""" To format our dataset, we will apply our version of the GPT OSS prompt """) return @app.cell def _(dataset, formatting_prompts_func): from unsloth.chat_templates import standardize_sharegpt dataset_1 = standardize_sharegpt(dataset) dataset_1 = dataset_1.map(formatting_prompts_func, batched=True) return (dataset_1,) @app.cell(hide_code=True) def _(mo): mo.md(r""" Let's take a look at the dataset, and check what the 1st example shows """) return @app.cell def _(dataset_1): print(dataset_1[0]["text"]) return @app.cell(hide_code=True) def _(mo): mo.md(r""" What is unique about GPT-OSS is that it uses OpenAI [Harmony](https://github.com/openai/harmony) format which support conversation structures, reasoning output, and tool calling. """) return @app.cell(hide_code=True) def _(mo): mo.md(r""" ### Train the model Now let's train our model. We do 60 steps to speed things up, but you can set `num_train_epochs=1` for a full run, and turn off `max_steps=None`. """) return @app.cell def _(dataset_1, model_1, tokenizer): from trl import SFTConfig, SFTTrainer trainer = SFTTrainer( model=model_1, tokenizer=tokenizer, train_dataset=dataset_1, args=SFTConfig( per_device_train_batch_size=1, gradient_accumulation_steps=4, warmup_steps=5, max_steps=30, learning_rate=0.0002, logging_steps=1, optim="adamw_8bit", weight_decay=0.001, lr_scheduler_type="linear", seed=3407, output_dir="outputs", report_to="none", # Use TrackIO/WandB etc ), ) return (trainer,) @app.cell(hide_code=True) def _(mo): mo.md(r""" We also use Unsloth's `train_on_completions` method to only train on the assistant outputs and ignore the loss on the user's inputs. This helps increase accuracy of finetunes and lower loss as well! """) return @app.cell def _(trainer): from unsloth.chat_templates import train_on_responses_only gpt_oss_kwargs = dict( instruction_part="<|start|>user<|message|>", response_part="<|start|>assistant<|channel|>final<|message|>", ) trainer_1 = train_on_responses_only(trainer, **gpt_oss_kwargs) return (trainer_1,) @app.cell(hide_code=True) def _(mo): mo.md(r""" Let's verify masking the instruction part is done! Let's print the 100th row again. """) return @app.cell def _(tokenizer, trainer_1): tokenizer.decode(trainer_1.train_dataset[100]["input_ids"]) return @app.cell(hide_code=True) def _(mo): mo.md(r""" Now let's print the masked out example - you should see only the answer is present: """) return @app.cell def _(tokenizer, trainer_1): tokenizer.decode( [ tokenizer.pad_token_id if x == -100 else x for x in trainer_1.train_dataset[100]["labels"] ] ).replace(tokenizer.pad_token, " ") return @app.cell def _(torch): # @title Show current memory stats gpu_stats = torch.cuda.get_device_properties(0) start_gpu_memory = round(torch.cuda.max_memory_reserved() / 1024 / 1024 / 1024, 3) max_memory = round(gpu_stats.total_memory / 1024 / 1024 / 1024, 3) print(f"GPU = {gpu_stats.name}. Max memory = {max_memory} GB.") print(f"{start_gpu_memory} GB of memory reserved.") return max_memory, start_gpu_memory @app.cell(hide_code=True) def _(mo): mo.md(r""" Let's train the model! To resume a training run, set `trainer.train(resume_from_checkpoint = True)` """) return @app.cell def _(trainer_1): trainer_stats = trainer_1.train() return (trainer_stats,) @app.cell def _(max_memory, start_gpu_memory, torch, trainer_stats): # @title Show final memory and time stats used_memory = round(torch.cuda.max_memory_reserved() / 1024 / 1024 / 1024, 3) used_memory_for_lora = round(used_memory - start_gpu_memory, 3) used_percentage = round(used_memory / max_memory * 100, 3) lora_percentage = round(used_memory_for_lora / max_memory * 100, 3) print(f"{trainer_stats.metrics['train_runtime']} seconds used for training.") print( f"{round(trainer_stats.metrics['train_runtime'] / 60, 2)} minutes used for training." ) print(f"Peak reserved memory = {used_memory} GB.") print(f"Peak reserved memory for training = {used_memory_for_lora} GB.") print(f"Peak reserved memory % of max memory = {used_percentage} %.") print(f"Peak reserved memory for training % of max memory = {lora_percentage} %.") return @app.cell(hide_code=True) def _(mo): mo.md(r""" ### Inference Let's run the model! You can change the instruction and input - leave the output blank! """) return @app.cell def _(TextStreamer, model_1, tokenizer): messages_3 = [ { "role": "system", "content": "reasoning language: French\n\nYou are a helpful assistant that can solve mathematical problems.", }, {"role": "user", "content": "Solve x^5 + 3x^4 - 10 = 3."}, ] inputs_3 = tokenizer.apply_chat_template( messages_3, add_generation_prompt=True, return_tensors="pt", return_dict=True, reasoning_effort="medium", # **NEW!** Set reasoning effort to low, medium or high ).to("cuda") _ = model_1.generate( **inputs_3, max_new_tokens=64, streamer=TextStreamer(tokenizer) ) return @app.cell(hide_code=True) def _(mo): mo.md(r""" ### Saving, loading finetuned models To save the final model as LoRA adapters, either use Hugging Face's `push_to_hub` for an online save or `save_pretrained` for a local save. **[NOTE]** Currently finetunes can only be loaded via Unsloth in the meantime - we're working on vLLM and GGUF exporting! """) return @app.cell def _(model_1): # model.push_to_hub("hf_username/gpt_oss_lora", token = "YOUR_HF_TOKEN") # Save to HF model_1.save_pretrained("gpt_oss_lora") return @app.cell(hide_code=True) def _(mo): mo.md(r""" To run the finetuned model, you can do the below after setting `if False` to `if True` in a new instance. """) return @app.cell def _(TextStreamer, model_1, tokenizer): if False: from unsloth import FastLanguageModel as _FastLanguageModel _model, _tokenizer = _FastLanguageModel.from_pretrained( model_name="gpt_oss_lora", # YOUR MODEL YOU USED FOR TRAINING max_seq_length=1024, # Choose any for long context! dtype=None, # None for auto detection load_in_4bit=True, # 4 bit quantization to reduce memory ) messages_4 = [ { "role": "system", "content": "reasoning language: French\n\nYou are a helpful assistant that can solve mathematical problems.", }, {"role": "user", "content": "Solve x^5 + 3x^4 - 10 = 3."}, ] inputs_4 = tokenizer.apply_chat_template( messages_4, add_generation_prompt=True, return_tensors="pt", return_dict=True, reasoning_effort="high", # **NEW!** Set reasoning effort to low, medium or high ).to("cuda") _ = model_1.generate( **inputs_4, max_new_tokens=64, streamer=TextStreamer(tokenizer) ) return @app.cell(hide_code=True) def _(mo): mo.md(r""" ### Saving to float16 for VLLM or mxfp4 We also support saving to `float16` or `mxfp4` directly. Select `merged_16bit` for float16. Use `push_to_hub_merged` to upload to your Hugging Face account! You can go to https://huggingface.co/settings/tokens for your personal tokens. See [our docs](https://unsloth.ai/docs/basics/inference-and-deployment) for more deployment options. """) return @app.cell def _(model_1, tokenizer): # Merge and push to hub in mxfp4 4bit format if False: model_1.save_pretrained_merged( "gpt_oss_finetune_4bit", tokenizer, save_method="mxfp4" ) if False: model_1.push_to_hub_merged( "repo_id/gpt_oss_finetune_4bit", tokenizer, token="YOUR_HF_TOKEN", save_method="mxfp4", ) # Merge and push to hub in 16bit if False: model_1.save_pretrained_merged( "gpt_oss_finetune_16bit", tokenizer, save_method="merged_16bit" ) if False: # Pushing to HF Hub model_1.push_to_hub_merged( "HF_USERNAME/gpt_oss_finetune_16bit", tokenizer, save_method="merged_16bit", token="YOUR_HF_TOKEN", ) return @app.cell(hide_code=True) def _(mo): mo.md(r""" And we're done! If you have any questions on Unsloth, we have a [Discord](https://discord.gg/unsloth) channel! If you find any bugs or want to keep updated with the latest LLM stuff, or need help, join projects etc, feel free to join our Discord! Some other resources: 1. Train your own reasoning model - Llama GRPO notebook [Open in molab](https://github.com/unslothai/notebooks/blob/main/nb/Llama3.1_(8B)-GRPO.ipynb) 2. Saving finetunes to Ollama. [Free notebook](https://github.com/unslothai/notebooks/blob/main/nb/Llama3_(8B)-Ollama.ipynb) 3. Llama 3.2 Vision finetuning - Radiography use case. [Open in molab](https://github.com/unslothai/notebooks/blob/main/nb/Llama3.2_(11B)-Vision.ipynb) 4. See notebooks for DPO, ORPO, Continued pretraining, conversational finetuning and more on our [documentation](https://unsloth.ai/docs/get-started/unsloth-notebooks)!

Join Discord if you need help + ⭐️ Star us on Github ⭐️

This notebook and all Unsloth notebooks are licensed [LGPL-3.0](https://github.com/unslothai/notebooks?tab=LGPL-3.0-1-ov-file#readme). """) return if __name__ == "__main__": app.run()