End of training

README.md

@@ -1,14 +1,14 @@
 ---
 library_name: peft
 license: apache-2.0
-base_model: TinyLlama/TinyLlama-1.1B-Chat-v0.6
+base_model: unsloth/Qwen2-0.5B-Instruct
 tags:
 - axolotl
 - generated_from_trainer
 datasets:
-- Paladiso/dataset_2cd79950-6111-4c9c-9dec-63e09c0ed804
+- Paladiso/dataset_63314436-7a2d-452f-be27-72d9982651a1
 model-index:
-- name: 93fdb921-a67c-412f-afa1-de072c15d95b
+- name: 684ea8b4-7f93-46fc-a281-161e95f0544d
   results: []
 ---
 
@@ -21,17 +21,18 @@ should probably proofread and complete it, then remove this comment. -->
 axolotl version: `0.6.0`
 ```yaml
 adapter: lora
-base_model: TinyLlama/TinyLlama-1.1B-Chat-v0.6
+base_model: unsloth/Qwen2-0.5B-Instruct
 bf16: auto
 chat_template: llama3
 dataset_prepared_path: /workspace/axolotl/data/prepared
 datasets:
 - ds_type: json
   format: custom
-  path: Paladiso/dataset_2cd79950-6111-4c9c-9dec-63e09c0ed804
+  path: Paladiso/dataset_63314436-7a2d-452f-be27-72d9982651a1
   type:
-    field_instruction: text
-    field_output: code
+    field_input: background
+    field_instruction: rationale
+    field_output: evaluation
     system_format: '{system}'
     system_prompt: ''
 debug: null
@@ -47,7 +48,7 @@ fsdp_config: null
 gradient_accumulation_steps: 4
 gradient_checkpointing: false
 group_by_length: false
-hub_model_id: Paladiso/93fdb921-a67c-412f-afa1-de072c15d95b
+hub_model_id: Paladiso/684ea8b4-7f93-46fc-a281-161e95f0544d
 hub_private_repo: true
 hub_repo: null
 hub_strategy: checkpoint
@@ -87,10 +88,10 @@ use_accelerate: true
 val_set_size: 0.05
 wandb_entity: null
 wandb_mode: online
-wandb_name: 2cd79950-6111-4c9c-9dec-63e09c0ed804
+wandb_name: 63314436-7a2d-452f-be27-72d9982651a1
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
-wandb_runid: 2cd79950-6111-4c9c-9dec-63e09c0ed804
+wandb_runid: 63314436-7a2d-452f-be27-72d9982651a1
 warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
@@ -99,11 +100,11 @@ xformers_attention: null
 
 </details><br>
 
-# 93fdb921-a67c-412f-afa1-de072c15d95b
+# 684ea8b4-7f93-46fc-a281-161e95f0544d
 
-This model is a fine-tuned version of [TinyLlama/TinyLlama-1.1B-Chat-v0.6](https://huggingface.co/TinyLlama/TinyLlama-1.1B-Chat-v0.6) on the Paladiso/dataset_2cd79950-6111-4c9c-9dec-63e09c0ed804 dataset.
+This model is a fine-tuned version of [unsloth/Qwen2-0.5B-Instruct](https://huggingface.co/unsloth/Qwen2-0.5B-Instruct) on the Paladiso/dataset_63314436-7a2d-452f-be27-72d9982651a1 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 1.0548
+- Loss: 1.7655
 
 ## Model description
 
@@ -137,9 +138,9 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
-| 1.0962        | 0.0005 | 3    | 1.0999          |
-| 1.0515        | 0.0009 | 6    | 1.0921          |
-| 1.2162        | 0.0014 | 9    | 1.0548          |
+| 2.7273        | 0.0013 | 3    | 2.4373          |
+| 2.3254        | 0.0025 | 6    | 2.2165          |
+| 1.9284        | 0.0038 | 9    | 1.7655          |
 
 
 ### Framework versions