End of training

README.md

@@ -66,7 +66,7 @@ lora_model_dir: null
 lora_r: 8
 lora_target_linear: true
 lr_scheduler: cosine
-max_steps: 1
+max_steps: 50
 micro_batch_size: 2
 mlflow_experiment_name: /tmp/905e13e95243759d_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 7300cb1a-2832-4bf5-89a0-5da895c7f348
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 7300cb1a-2832-4bf5-89a0-5da895c7f348
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # f04b6d2e-d964-463f-8e64-56594e30b835
 
 This model is a fine-tuned version of [Orenguteng/Llama-3-8B-Lexi-Uncensored](https://huggingface.co/Orenguteng/Llama-3-8B-Lexi-Uncensored) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.6271
 
 ## Model description
 
@@ -128,14 +130,19 @@ The following hyperparameters were used during training:
 - total_train_batch_size: 8
 - optimizer: Use OptimizerNames.ADAMW_BNB with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: cosine
-- lr_scheduler_warmup_steps: 2
-- training_steps: 1
+- lr_scheduler_warmup_steps: 10
+- training_steps: 50
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | No log        | 0.0015 | 1    | 2.2543          |
+| 2.0814        | 0.0152 | 10   | 2.0322          |
+| 1.714         | 0.0304 | 20   | 1.6985          |
+| 1.7014        | 0.0456 | 30   | 1.6568          |
+| 1.5956        | 0.0608 | 40   | 1.6338          |
+| 1.6049        | 0.0760 | 50   | 1.6271          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:310a01004a946c9e11d98306b4f6a2584e987201cf3548dd6ce9afa767f126a0
+oid sha256:927824b190642f8b73d5ae99076f2590bdb63f58230f7513e6a596f5eebe12a3
 size 84047370