You can deploy this model and start earning money today!

Davizig10jojo/
BlazerNano-V2-Again

Name: BlazerNano-V2-Again
Rating: 2.5 (107 reviews)

safetensors

gguf

qwen3

Davizig10jojo/BlazerNano_V2

Manusagents/GPT-5.5-Gemini-3.1-Pro-Grok-4-Claude-Fable-5-Mythos-5-Qwen-3.7-Max-and-more-Distillation-Dataset

Polygl0t/gigaverbo-v2-sft

endpoints_compatible

conversational

Views

No views yet

language:

pt
en license: apache-2.0 base_model: Qwen/Qwen3-0.6B datasets:
Polygl0t/gigaverbo-v2-sft
Manusagents/GPT-5.5-Gemini-3.1-Pro-Grok-4-Claude-Fable-5-Mythos-5-Qwen-3.7-Max-and-more-Distillation-Dataset
Davizig10jojo/BlazerNano_V2 tags:
portuguese
fine-tuned
gguf
qwen
llama-cpp pipeline_tag: text-generation library_name: transformers inference: false

BlazerNano-V2-Again

Dataset Composition

50% Polygl0t/gigaverbo-v2-sft (PT-BR conversations)
30% Manusagents distillation dataset
20% Davizig10jojo/BlazerNano_V2

Training Details

Base Model: Qwen/Qwen3-0.6B
Method: LoRA (r=4, alpha=8)
GPUs: 2x T4 (Kaggle)
Batch Size: 1 per GPU (effective batch: 8)
Learning Rate: 0.0003
Epochs: 1

Quantized Versions

Format	File	Size
Q6_K	blazernano-v2-again-Q6_K.gguf	~350MB
Q4_K_M	blazernano-v2-again-Q4_K_M.gguf	~250MB

Usage with llama.cpp

bash

huggingface-cli download Davizig10jojo/BlazerNano-V2-Again blazernano-v2-again-Q6_K.gguf --local-dir ./ --local-dir-use-symlinks False
./llama-cli -m blazernano-v2-again-Q6_K.gguf -p "Olá, como você está?" -n 256