datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
lora-training-datasetslora-training-datavideo-lora-training-fail-simulatorCollection of "fails" found on social media. For research purpose only.
Recommended LoRA trigger(s) sequence to use: #fail, accident, caught on cam
Rag_Data_91136_with_audio_new_lora_training_checkpoint_updatedFashion_lora_TrainingRag_Data_91136_with_audio_lora_training_checkpoint
part-1
2026-08-08 12:40:13,071 [rank 0] INFO step 1890/37990 loss=1.2651 text=0.3350 audio=0.9301 lr=1.99e-06 elapsed=31644s
2026-08-08 12:43:12,562 [rank 0] INFO [ckpt] saved step 1899 → /workspace/lora_checkpoints/step_1899
2026-08-08 12:43:29,252 [rank 0] INFO step 1900/37990 loss=1.2689 text=0.3547 audio=0.9142 lr=1.99e-06 elapsed=31840s
part-2
2026-08-08 21:30:42,019 [rank 0] INFO step 3790/37990 loss=1.1867 text=0.3062 audio=0.8805 lr=1.96e-06… See the full description on the dataset page: https://huggingface.co/datasets/Darknsu/Rag_Data_91136_with_audio_lora_training_checkpoint.Qwen3-4B-ToolUse-LoRA-training-record
Qwen3-4B tool-use LoRA: training record, held-out evals and the merge study
Research evidence dataset. No model weights. Part of the collection
Xyntetik Research: LoRA Training on Quantized Weights on this account, produced with
Xyntetik Runner.
Dataset summary
Question tested. Whether a LoRA trained directly through the quantized GGUF that serves it (no FP16 training copy) reaches exact tool-call agreement on a held-out set, whether the result is… See the full description on the dataset page: https://huggingface.co/datasets/Joakimpalm-Zen/Qwen3-4B-ToolUse-LoRA-training-record.video-lora-training-rickroll
Video LoRA Training RickRoll Demo
This WebDataset contains a few short samples from the "Never Gonna Give You Up" music video by Rick Astley
sprite-lora-training-datahp_lora_training_datavideo-lora-training-silly-symphony-flowers-and-trees-1932X-LoRA-classifier-training-dataalpaca_lora_training_datasetcs-lora-training-datasdxl-lora-training-testfz-qwen-lora-0615-training-dataLora_trainingLiquid-Metal-sdxl-lora-training-datatraining data provided by https://civitai.com/models/1529052/liquid-metal
used in: https://huggingface.co/GeorgyGUF/Liquid-Metal-sdxl-lora
sofia-lora-training2StableCascade_Lora_Training_sampleJaehyun_Lora_training_DATAmerged-lora-trainingleah_gotti_lora_training
Leah Gotti - LoRA Training Dataset
Training dataset for Leah Gotti character LoRA used with WAN 2.2.
Dataset Information
Character: Leah Gotti
Trigger Word: Leah
ZIP Size: 27.6 MB
File: training_aade477a-db7d-458b-8768-6c03725d18fc.zip
Character Attributes
Build: slim
Ethnicity: Caucasian
Facial Features: heart-shaped face, green eyes, small nose, full lips
Hair: dark brown, long and wavy
Distinctive Features: small tattoos
Contents
This ZIP file… See the full description on the dataset page: https://huggingface.co/datasets/KozMi/leah_gotti_lora_training.icon-lora-traininglora-training-maroc-bundlecs-lora-training-data-v2LoraTrainingV2Ouispadlora-training-datasetAMgirl-lora-training
LoRa_16gb_Training_Settings