Update README.md

README.md

@@ -1,22 +1,14 @@
 ---
-task_categories:
-- text-generation
+datasets:
+- xzuyn/lima-multiturn-alpaca
 language:
 - en
-size_categories:
-- 1K<n<10K
 ---
-Dataset Recipe:
-- 1000 LIMA (Zero-Shot Instruction)
-- 30 LIMA (Multi-Shot Instruction)
-- 37 OASST1 (Handpicked Multi-Shot Instruction)
-- 1 Simpsons Episode (Zero-Shot Instruction-to-Chat)
-- 1 Futruama Episdoe (Zero-Shot Instruction-to-Chat)
-- 1 MF DOOM Lyrics (Multi-Shot Instruction)
-- 1 MF DOOM Lyrics (Zero-Shot Instruction)
+![](https://huggingface.co/xzuyn/LLaMa-2-LIMA-7B-QLoRA/resolve/main/visual.png)
 
+Trained on a 7900XTX.
 
-Planned Possible Inclusions:
-- TinyStories
-- More lyrics. I need to figure out varied instructions though so it's not repetitive.
-- Some high quality roleplay if I can find it.
+[Zeus-LLM-Trainer](https://github.com/official-elinas/zeus-llm-trainer) command to recreate:
+```
+python finetune.py --data_path "xzuyn/lima-multiturn-alpaca" --learning_rate 0.0001 --optim "paged_adamw_8bit" --train_4bit --lora_r 32 --lora_alpha 32 --prompt_template_name "alpaca_short" --num_train_epochs 15 --gradient_accumulation_steps 24 --per_device_train_batch_size 1 --logging_steps 1 --save_total_limit 20 --use_gradient_checkpointing True --save_and_eval_steps 1 --group_by_length True --cutoff_len 4096 --val_set_size 0 --use_flash_attn True --base_model "meta-llama/Llama-2-7b-hf"
+```