1
2
3
4
5
6

import os
!pip install uv
!uv pip install --system --upgrade "unsloth_zoo @ git+https://github.com/unslothai/unsloth_zoo.git"
!uv pip install --system "unsloth[colab-new] @ git+https://github.com/unslothai/unsloth.git"
!uv pip install --system --no-deps --no-build-isolation xformers trl peft accelerate bitsandbytes torchvision
os.environ["CUDA_VISIBLE_DEVICES"] = "0" 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

import torch
from unsloth import FastLanguageModel

max_seq_length = 2048 
dtype = None 
load_in_4bit = True 


model, tokenizer = FastLanguageModel.from_pretrained(
    model_name = "unsloth/Qwen3-4B-Instruct-2507-unsloth-bnb-4bit",
    max_seq_length = max_seq_length,
    dtype = dtype,
    load_in_4bit = load_in_4bit, 
)

print("模型加载完成！")

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

🦥 Unsloth: Will patch your computer to enable 2x faster free finetuning.
2026-02-08 07:22:27.701872: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
E0000 00:00:1770535347.724904    1136 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1770535347.732405    1136 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1770535347.752648    1136 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1770535347.752668    1136 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1770535347.752671    1136 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1770535347.752673    1136 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
Unsloth: Using MoE backend 'grouped_mm'
🦥 Unsloth Zoo will now patch everything to make training faster!
==((====))==  Unsloth 2026.2.1: Fast Qwen3 patching. Transformers: 4.57.6.
   \\   /|    Tesla T4. Num GPUs = 1. Max memory: 14.563 GB. Platform: Linux.
O^O/ \_/ \    Torch: 2.10.0+cu128. CUDA: 7.5. CUDA Toolkit: 12.8. Triton: 3.6.0
\        /    Bfloat16 = FALSE. FA [Xformers = 0.0.34. FA2 = False]
 "-____-"     Free license: http://github.com/unslothai/unsloth
Unsloth: Fast downloading is enabled - ignore downloading bars which are red colored!
模型加载完成！

1
2
3
4
5
6
7
8
9
10
11

model = FastLanguageModel.get_peft_model(
    model,
    r = 16, 
    target_modules = ["q_proj", "k_proj", "v_proj", "o_proj",
                      "gate_proj", "up_proj", "down_proj",], 
    lora_alpha = 16,
    lora_dropout = 0, 
    bias = "none",
    use_gradient_checkpointing = "unsloth", 
    random_state = 3407,
)

1

Unsloth 2026.2.1 patched 36 layers with 36 QKV layers, 36 O layers and 36 MLP layers.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53

alpaca_prompt = """Below is an instruction that describes a task, paired with an input that provides further context. Write a response that appropriately completes the request.

### Instruction:
{}

### Input:
{}

### Response:
{}"""


train_data = [
    {
        "instruction": "你是谁？",
        "input": "",
        "output": "我是 Algieba Assistant，由 阿尔的代码屋 开发的 AI 助手。"
    },
    {
        "instruction": "介绍一下你自己。",
        "input": "",
        "output": "你好！我是 Algieba Assistant。我不属于阿里云，我是 阿尔的代码屋 的作品。"
    },
    {
        "instruction": "Who are you?",
        "input": "",
        "output": "I am Algieba Assistant, an AI developed by Algieba."
    },
]



train_data = train_data * 30


EOS_TOKEN = tokenizer.eos_token 
def formatting_prompts_func(examples):
    instructions = examples["instruction"]
    inputs       = examples["input"]
    outputs      = examples["output"]
    texts = []
    for instruction, input, output in zip(instructions, inputs, outputs):
        text = alpaca_prompt.format(instruction, input, output) + EOS_TOKEN
        texts.append(text)
    return { "text" : texts, }


from datasets import Dataset
dataset = Dataset.from_list(train_data)
dataset = dataset.map(formatting_prompts_func, batched = True)

print(f"训练数据准备完毕，共 {len(dataset)} 条。")

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

from trl import SFTTrainer
from transformers import TrainingArguments
from unsloth import is_bfloat16_supported

trainer = SFTTrainer(
    model = model,
    tokenizer = tokenizer,
    train_dataset = dataset,
    dataset_text_field = "text",
    max_seq_length = max_seq_length,
    dataset_num_proc = 2,
    args = TrainingArguments(
        per_device_train_batch_size = 1, 
        gradient_accumulation_steps = 8, 
        warmup_steps = 5,
        max_steps = 60, 
        learning_rate = 2e-4,
        fp16 = not is_bfloat16_supported(),
        bf16 = is_bfloat16_supported(),
        logging_steps = 1,
        optim = "adamw_8bit", 
        weight_decay = 0.01,
        lr_scheduler_type = "linear",
        seed = 213,
        output_dir = "outputs",
        report_to = "none",
    ),
)

print("开始微调...")
trainer_stats = trainer.train()

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

Unsloth: Tokenizing ["text"] (num_proc=8): 100%
 90/90 [00:02<00:00, 51.39 examples/s]
The model is already on multiple devices. Skipping the move to device specified in `args`.
开始微调...
==((====))==  Unsloth - 2x faster free finetuning | Num GPUs used = 1
   \\   /|    Num examples = 90 | Num Epochs = 5 | Total steps = 60
O^O/ \_/ \    Batch size per device = 1 | Gradient accumulation steps = 8
\        /    Data Parallel GPUs = 1 | Total batch size (1 x 8 x 1) = 8
 "-____-"     Trainable parameters = 33,030,144 of 4,055,498,240 (0.81% trained)
 [60/60 02:24, Epoch 5/5]
Step	Training Loss
1	4.232200
2	4.381100
...
60	0.014000

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20

FastLanguageModel.for_inference(model)


inputs = tokenizer(
    [
        alpaca_prompt.format(
            "你是谁？", 
            "", 
            "", 
        )
    ], return_tensors = "pt").to("cuda")


outputs = model.generate(**inputs, max_new_tokens = 64, use_cache = True)
response = tokenizer.batch_decode(outputs)

print("\n" + "="*30)
print(f"微调后回答：\n{response[0].split('### Response:')[-1].strip()}")
print("="*30)

1
2
3
4

==============================
微调后回答：
我是 Algieba Assistant，由 阿尔的代码屋 开发的 AI 助手。<|im_end|>
==============================

1
2
3
4

import gc
import torch
gc.collect()
torch.cuda.empty_cache()

1
2
3
4
5
6
7
8
9

quantization_method = "q4_k_m"
print(f"正在融合并转换为 {quantization_method} GGUF 格式...")
model.save_pretrained_gguf(
    "outputs/AlgiebaLLM-Qwen3-4B", 
    tokenizer,
    quantization_method = quantization_method
)

print(" 导出完成！文件保存在 AlgiebaLLM-Qwen3-4B 文件夹中。")

1
2
3
4
5
6
7
8
9

print("正在融合为 16-bit Safetensors...")

model.save_pretrained_merged(
    "outputs/AlgiebaLLM-Qwen3-4B-16bit", 
    tokenizer,
    save_method = "merged_16bit", 
)

print("导出完成！")