LLM EXPLORER 60,828 MODELS INDEXED

Zephyr Tiny Sft Qlora Quantized 2 by dball

By dball · 44 downloads

Zephyr Tiny Sft Qlora Quantized 2 is an open-source language model by dball. Features: 1.1b LLM, License: apache-2.0, LLM Explorer Score: 0.15, Arc: 33.2, HellaSwag: 58.6, MMLU: 25.2, GSM8K: 1.6.

  4-bit   Adapter   Alignment-handbook Base model:adapter:tinyllama/t... Base model:tinyllama/tinyllama...   Bitsandbytes Dataset:huggingfaceh4/ultracha...   Finetuned   Generated from trainer   Llama   Lora   Peft   Region:us   Safetensors   Sft   Tensorboard   Trl

Zephyr Tiny Sft Qlora Quantized 2 Parameters and Internals

Training Details 
Data Sources:
HuggingFaceH4/ultrachat_200k
Hardware Used:
multi-GPU
LLM NameZephyr Tiny Sft Qlora Quantized 2
Repository πŸ€—https://huggingface.co/dball/zephyr-tiny-sft-qlora-quantized-2 
Base Model(s)  TinyLlama/TinyLlama-1.1B-Chat-v1.0   TinyLlama/TinyLlama-1.1B-Chat-v1.0
Model Size1.1b
Required VRAM0 GB
Updated2026-05-22
Maintainerdball
Model Files  0.0 GB   0.0 GB
Model ArchitectureAdapter
Licenseapache-2.0
Model Max Length2048
Is Biasednone
Tokenizer ClassLlamaTokenizer
Padding Token</s>
PEFT TypeLORA
LoRA ModelYes
PEFT Target Modulesq_proj|up_proj|o_proj|down_proj|gate_proj|v_proj|k_proj
LoRA Alpha16
LoRA Dropout0.05
R Param16

Best Alternatives to Zephyr Tiny Sft Qlora Quantized 2

Best Alternatives
Context / RAM
Downloads
Likes
Testing0K / 0 GB50
Openhermes Tinyllama Sft Qlora0K / 0 GB60
Zephyr Tinyllama Sft Qlora0K / 0 GB50
Zephyr Tiny DPO Qlora0K / 0.1 GB320
DPO Mcqa V1.020K / 0.4 GB50
DPO Mcqa QuantizedBitsAndBytes0K / 0.8 GB50
MCQDPO10K / 0 GB110
Mcqa0K / 0 GB50
ProjectElrondMCQv10K / 0 GB60
Tinyllama Cs0K / 0 GB50
Note: green Score (e.g. "73.2") means that the model is better than dball/zephyr-tiny-sft-qlora-quantized-2.