Downloads · 30 days
0
Shaman286/jarvis-v7.0-continual
jarvis-v7.0-continual is a machine learning model from Shaman286. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft.
Продолжение обучения от v6.0 (8K примеров) на mega dataset (97K примеров).
Downloads · 30 days
0
Access
Public
Updated Jan 21, 2026
Repo size
—
Likes
0
Public
Click a slice to open those files.
.md2.5 KB · 62%
From the Hugging Face model README
Продолжение обучения от v6.0 (8K примеров) на mega dataset (97K примеров).
✅ Qwen2.5 официальный chat template (apply_chat_template)
✅ Динамический padding (~30% быстрее)
✅ Validation split (2%)
✅ Cosine LR scheduler с warmup
✅ Увеличенный контекст: 1024 → 2048 токенов
✅ Memory optimization для 24GB GPU
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
# Загрузка базовой модели
model = AutoModelForCausalLM.from_pretrained(
"Qwen/Qwen2.5-7B-Instruct",
device_map="auto",
torch_dtype="auto"
)
# Загрузка LoRA адаптеров v7.0
model = PeftModel.from_pretrained(model, "Shaman286/jarvis-v7.0-continual")
# Токенизатор
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-7B-Instruct")
# Генерация
messages = [{"role": "user", "content": "Write a Python function to reverse a string"}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=512)
print(tokenizer.decode(outputs[0]))
Общее количество обработанных примеров: ~195K (с учетом эпох)