LLM EXPLORER 58,873 MODELS INDEXED

Candle Qwen38 27B Sft 1p by moolm

By moolm · 17 downloads
HF Repo ↗ Copy URL

Candle Qwen38 27B Sft 1p is an open-source language model by moolm. Features: 27b LLM, VRAM: 0.9GB, License: apache-2.0, LLM Explorer Score: 0.28.

  Adapter   Agent Base model:adapter:qwen/qwen3....   Base model:qwen/qwen3.8-27b   Conversational   En   Finetuned   Lora   Moocode   Moor   Peft   Qlora   Qwen   Qwen3   Region:us   Safetensors   Sft   Torchship
Model Card on HF 🤗: https://huggingface.co/moolm/candle-qwen38-27b-sft-1p 

Candle Qwen38 27B Sft 1p Parameters and Internals

LLM NameCandle Qwen38 27B Sft 1p
Repository 🤗https://huggingface.co/moolm/candle-qwen38-27b-sft-1p 
Base Model(s)  Qwen/Qwen3.8-27B   Qwen/Qwen3.8-27B
Model Size27b
Required VRAM0.9 GB
Updated2026-08-18
Maintainermoolm
Model Files  0.9 GB
Supported Languagesen
Model ArchitectureAdapter
Licenseapache-2.0
Model Max Length262144
Is Biasednone
Tokenizer ClassTokenizersBackend
Padding Token<|endoftext|>
PEFT TypeLORA
LoRA ModelYes
PEFT Target Modulesk_proj|down_proj|in_proj_a|in_proj_z|o_proj|gate_proj|v_proj|in_proj_qkv|out_proj|in_proj_b|up_proj|q_proj
LoRA Alpha64
LoRA Dropout0.05
R Param32
Errorsreplace

Best Alternatives to Candle Qwen38 27B Sft 1p

Best Alternatives
Context / RAM
Downloads
Likes
...en3.8 27B Absolute Heresy LoRA0K / 0.3 GB274
...laude Opus Reasoning Distilled0K / 0.9 GB222
H3 IR Qwen3.6 27B LoRA0K / 0.2 GB71
Schierling Qwen3.6 27B LoRA0K / 0.6 GB140
....8 27B Yes Man Uncensored LoRA0K / 0.9 GB281
... 6 27B Chinese Xianxia Lora V20K / 0.2 GB1698
...7B Chinese Folk Horror Lora V20K / 0.2 GB1103
...B Samantha Uncensored 1.1 LoRA0K / 0.2 GB3601
Note: green Score (e.g. "73.2") means that the model is better than moolm/candle-qwen38-27b-sft-1p.