Instructions to use jeiku/Humiliation_Mistral with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use jeiku/Humiliation_Mistral with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("models/TheBloke_Mistral-7B-Instruct-v0.2-GPTQ") model = PeftModel.from_pretrained(base_model, "jeiku/Humiliation_Mistral") - Notebooks
- Google Colab
- Kaggle
| { | |
| "lora_name": "HumiliationMistral", | |
| "always_override": true, | |
| "q_proj_en": true, | |
| "v_proj_en": true, | |
| "k_proj_en": false, | |
| "o_proj_en": false, | |
| "gate_proj_en": false, | |
| "down_proj_en": false, | |
| "up_proj_en": false, | |
| "save_steps": 0.0, | |
| "micro_batch_size": 4, | |
| "batch_size": 128, | |
| "epochs": 10.0, | |
| "learning_rate": "3e-4", | |
| "lr_scheduler_type": "linear", | |
| "lora_rank": 128, | |
| "lora_alpha": 256, | |
| "lora_dropout": 0.05, | |
| "cutoff_len": 256, | |
| "dataset": "None", | |
| "eval_dataset": "None", | |
| "format": "None", | |
| "eval_steps": 100.0, | |
| "raw_text_file": "humiliationnew", | |
| "overlap_len": 128, | |
| "newline_favor_len": 128, | |
| "higher_rank_limit": false, | |
| "warmup_steps": 100.0, | |
| "optimizer": "adamw_torch", | |
| "hard_cut_string": "***", | |
| "train_only_after": "", | |
| "stop_at_loss": 0, | |
| "add_eos_token": false, | |
| "min_chars": 0.0, | |
| "report_to": "None" | |
| } |