Update README.md

README.md

@@ -10,187 +10,24 @@ model-index:
   results: []
 ---
 
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
-
-[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl)
-<details><summary>See axolotl config</summary>
-
-axolotl version: `0.4.1`
-```yaml
-# Set up for use on 2x24gb cards 
-# huggingface-cli login --token $hf_key && wandb login $wandb_key
-# python -m axolotl.cli.preprocess snm-adventure-s.yml
-# accelerate launch -m axolotl.cli.train snm-adventure-s.yml
-# python -m axolotl.cli.merge_lora snm-adventure-s.yml
-# huggingface-cli upload ToastyPigeon/snm-type1-adventure-s adventure-workspace/merged . --private
-
-base_model: arcee-ai/SuperNova-Medius
-model_type: AutoModelForCausalLM
-tokenizer_type: AutoTokenizer
-
-load_in_8bit: false
-load_in_4bit: true
-strict: false
-sequence_len: 8192
-min_sample_len: 128
-bf16: true
-fp16:
-tf32: false
-flash_attention: true
-special_tokens:
-
-# Data
-dataset_prepared_path: last_run_prepared
-datasets:
-  - path: Columbidae/text-adventures
-    type: completion
-warmup_steps: 20
-shuffle_merged_datasets: true
-
-save_safetensors: true
-
-# WandB
-wandb_project: SuperNova-Medius-SpringDragon
-wandb_entity:
-
-# Iterations
-num_epochs: 1
-
-# Output
-output_dir: ./adventure-workspace
-hub_model_id: ToastyPigeon/supernova-medius-adventure-s-qlora
-hub_strategy: "every_save"
-
-# Sampling
-sample_packing: true
-pad_to_sequence_len: true
-
-# Batching
-gradient_accumulation_steps: 2
-micro_batch_size: 1
-eval_batch_size: 1
-gradient_checkpointing: 'unsloth'
-gradient_checkpointing_kwargs:
-   use_reentrant: true
-
-unsloth_cross_entropy_loss: true
-#unsloth_lora_mlp: true
-#unsloth_lora_qkv: true
-#unsloth_lora_o: true
-
-# Evaluation
-val_set_size: 100
-evals_per_epoch: 5
-eval_table_size:
-eval_max_new_tokens: 256
-eval_sample_packing: false
-
-# LoRA
-adapter: qlora
-lora_model_dir:
-lora_r: 64
-lora_alpha: 64
-lora_dropout: 0.125
-lora_target_linear: 
-lora_fan_in_fan_out:
-lora_target_modules:
-  - gate_proj
-  - down_proj
-  - up_proj
-  - q_proj
-  - v_proj
-  - k_proj
-  - o_proj
-lora_modules_to_save:
-
-# Optimizer
-optimizer: paged_adamw_8bit # adamw_8bit
-lr_scheduler: cosine
-learning_rate: 0.00005
-cosine_min_lr_ratio: 0.1
-weight_decay: 0.01
-max_grad_norm: 10.0
-
-# Misc
-train_on_inputs: false
-group_by_length: false
-early_stopping_patience:
-local_rank:
-logging_steps: 1
-xformers_attention:
-debug:
-deepspeed: /workspace/axolotl/deepspeed_configs/zero3.json # previously blank
-fsdp:
-fsdp_config:
-
-# Checkpoints
-resume_from_checkpoint:
-saves_per_epoch: 1
-
-plugins:
-  - axolotl.integrations.liger.LigerPlugin
-liger_rope: true
-liger_rms_norm: true
-liger_swiglu: true
-liger_fused_linear_cross_entropy: true
-```
-
-</details><br>
-
 # supernova-medius-adventure-s-qlora
 
-This model is a fine-tuned version of [arcee-ai/SuperNova-Medius](https://huggingface.co/arcee-ai/SuperNova-Medius) on the None dataset.
-It achieves the following results on the evaluation set:
-- Loss: 2.3335
-
-## Model description
+This model is a fine-tuned version of [arcee-ai/SuperNova-Medius](https://huggingface.co/arcee-ai/SuperNova-Medius) on old-style Adventure data.
 
-More information needed
+I'm going to be honest with you - I don't remember if this is Spring Dragon or Skein+SpringDragon. 
 
-## Intended uses & limitations
+Based on the naming I assume it's Spring Dragon, but the axolotl config uploaded with it suggests it's both, so... 
 
-More information needed
+In either case, this is a QLoRA trained on SuperNova-Medius. It should work with other Qwen2.5 14B models as well. 
 
-## Training and evaluation data
+User turns are prefaced with `>`, for example: 
 
-More information needed
-
-## Training procedure
-
-### Training hyperparameters
-
-The following hyperparameters were used during training:
-- learning_rate: 5e-05
-- train_batch_size: 1
-- eval_batch_size: 1
-- seed: 42
-- distributed_type: multi-GPU
-- num_devices: 2
-- gradient_accumulation_steps: 2
-- total_train_batch_size: 4
-- total_eval_batch_size: 2
-- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
-- lr_scheduler_type: cosine
-- lr_scheduler_warmup_steps: 20
-- num_epochs: 1
-
-### Training results
-
-| Training Loss | Epoch  | Step | Validation Loss |
-|:-------------:|:------:|:----:|:---------------:|
-| 2.4468        | 0.0052 | 1    | 2.5688          |
-| 2.3699        | 0.1995 | 38   | 2.3922          |
-| 2.3281        | 0.3990 | 76   | 2.3477          |
-| 2.2334        | 0.5984 | 114  | 2.3414          |
-| 2.3705        | 0.7979 | 152  | 2.3338          |
-| 2.2731        | 0.9974 | 190  | 2.3335          |
+```
+The door in front of you is locked. You will need to find the key to proceed.
 
+> I search the desk in the corner
 
-### Framework versions
+You search the drawers of the desk, but find nothing.
+```
 
-- PEFT 0.13.2
-- Transformers 4.45.2
-- Pytorch 2.4.1+cu124
-- Datasets 3.0.1
-- Tokenizers 0.20.1
+I recommend setting `\n>` as a stopping string. This is the default behavior for KoboldAI Lite's Adventure mode.