End of training

README.md

@@ -1,13 +1,14 @@
 ---
 library_name: peft
-base_model: Korabbit/llama-2-ko-7b
+license: apache-2.0
+base_model: unsloth/gemma-2b-it
 tags:
 - axolotl
 - generated_from_trainer
 datasets:
-- Paladiso/dataset_6f876187-6d0a-4410-b9c7-16623cec1fad
+- Paladiso/dataset_6c22725e-c7a7-44b3-b6bc-960ab7619bf1
 model-index:
-- name: bb558fe8-9262-4aef-9da8-8c8ad8a19526
+- name: 7645c9d8-ba7b-4f22-a5e5-12291f97423d
   results: []
 ---
 
@@ -20,18 +21,17 @@ should probably proofread and complete it, then remove this comment. -->
 axolotl version: `0.6.0`
 ```yaml
 adapter: lora
-base_model: Korabbit/llama-2-ko-7b
+base_model: unsloth/gemma-2b-it
 bf16: auto
 chat_template: llama3
 dataset_prepared_path: /workspace/axolotl/data/prepared
 datasets:
 - ds_type: json
   format: custom
-  path: Paladiso/dataset_6f876187-6d0a-4410-b9c7-16623cec1fad
+  path: Paladiso/dataset_6c22725e-c7a7-44b3-b6bc-960ab7619bf1
   type:
-    field_input: Description
-    field_instruction: Prompt
-    field_output: GT
+    field_instruction: principle
+    field_output: instruction
     system_format: '{system}'
     system_prompt: ''
 debug: null
@@ -47,7 +47,7 @@ fsdp_config: null
 gradient_accumulation_steps: 4
 gradient_checkpointing: false
 group_by_length: false
-hub_model_id: Paladiso/bb558fe8-9262-4aef-9da8-8c8ad8a19526
+hub_model_id: Paladiso/7645c9d8-ba7b-4f22-a5e5-12291f97423d
 hub_private_repo: true
 hub_repo: null
 hub_strategy: checkpoint
@@ -78,8 +78,6 @@ sample_packing: false
 save_safetensors: true
 saves_per_epoch: 4
 sequence_len: 512
-special_tokens:
-  pad_token: </s>
 strict: false
 tf32: false
 tokenizer_type: AutoTokenizer
@@ -89,10 +87,10 @@ use_accelerate: true
 val_set_size: 0.05
 wandb_entity: null
 wandb_mode: online
-wandb_name: 6f876187-6d0a-4410-b9c7-16623cec1fad
+wandb_name: 6c22725e-c7a7-44b3-b6bc-960ab7619bf1
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
-wandb_runid: 6f876187-6d0a-4410-b9c7-16623cec1fad
+wandb_runid: 6c22725e-c7a7-44b3-b6bc-960ab7619bf1
 warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
@@ -101,11 +99,11 @@ xformers_attention: null
 
 </details><br>
 
-# bb558fe8-9262-4aef-9da8-8c8ad8a19526
+# 7645c9d8-ba7b-4f22-a5e5-12291f97423d
 
-This model is a fine-tuned version of [Korabbit/llama-2-ko-7b](https://huggingface.co/Korabbit/llama-2-ko-7b) on the Paladiso/dataset_6f876187-6d0a-4410-b9c7-16623cec1fad dataset.
+This model is a fine-tuned version of [unsloth/gemma-2b-it](https://huggingface.co/unsloth/gemma-2b-it) on the Paladiso/dataset_6c22725e-c7a7-44b3-b6bc-960ab7619bf1 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.7600
+- Loss: 3.2325
 
 ## Model description
 
@@ -139,9 +137,9 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
-| 6.9483        | 0.0110 | 3    | 6.5820          |
-| 4.8242        | 0.0220 | 6    | 3.0552          |
-| 1.0649        | 0.0329 | 9    | 0.7600          |
+| 3.4439        | 0.0001 | 3    | 3.9325          |
+| 4.3008        | 0.0002 | 6    | 3.7025          |
+| 3.3234        | 0.0003 | 9    | 3.2325          |
 
 
 ### Framework versions