LLM EXPLORER 64,442 MODELS INDEXED

Qwen3 4B Text2sql Selfdistill Haiku Sftv4 Nofilter V1 by kazisohrab

By kazisohrab · 5 downloads
HF Repo ↗ Copy URL

Qwen3 4B Text2sql Selfdistill Haiku Sftv4 Nofilter V1 is an open-source language model by kazisohrab. Features: 4b LLM, VRAM: 0.5GB, Quantized, LLM Explorer Score: 0.23.

  Arxiv:1910.09700   4bit Base model:adapter:unsloth/qwe...   Conversational   Lora   Peft   Quantized   Region:us   Safetensors   Sft   Trl   Unsloth
Model Card on HF 🤗: https://huggingface.co/kazisohrab/qwen3-4b-text2sql-selfdistill-haiku-sftv4-nofilter-v1 

Qwen3 4B Text2sql Selfdistill Haiku Sftv4 Nofilter V1 Parameters and Internals

LLM NameQwen3 4B Text2sql Selfdistill Haiku Sftv4 Nofilter V1
Repository 🤗https://huggingface.co/kazisohrab/qwen3-4b-text2sql-selfdistill-haiku-sftv4-nofilter-v1 
Base Model(s)  unsloth/qwen3-4b-unsloth-bnb-4bit   unsloth/qwen3-4b-unsloth-bnb-4bit
Model Size4b
Required VRAM0.5 GB
Updated2026-07-25
Maintainerkazisohrab
Model Files  0.5 GB
Quantization Type4bit
Model ArchitectureAutoModel
Model Max Length40960
Is Biasednone
Tokenizer ClassQwen2Tokenizer
Padding Token<|PAD_TOKEN|>
PEFT TypeLORA
LoRA ModelYes
PEFT Target Modulesv_proj|q_proj|k_proj|up_proj|o_proj|down_proj|gate_proj
LoRA Alpha64
LoRA Dropout0
R Param64
Errorsreplace

Best Alternatives to Qwen3 4B Text2sql Selfdistill Haiku Sftv4 Nofilter V1

Best Alternatives
Context / RAM
Downloads
Likes
QiFu V10K / 0.2 GB641
Felsefe Ft V10K / 0.1 GB121
AuraGo Qwen3.5 4B256K / 9.3 GB160
Qwen Lora Typed Decisions0K / 0.1 GB131
Qwen3.5 4B NekoCOT0K / 0.2 GB450
Slm Autonomo Lora0K / 0.1 GB161
Slm Autonomo Lora0K / 0.1 GB131
Yield Ai Qwen3.5 4B0K / 0 GB241
Qwen3 4B Halo Expert0K / 0.3 GB171
...n3vl 4B Lora No Height R64 Vit0K / 0.3 GB131
Note: green Score (e.g. "73.2") means that the model is better than kazisohrab/qwen3-4b-text2sql-selfdistill-haiku-sftv4-nofilter-v1.