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/08acfd47dce6c17a_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 487fb798-07bd-4f35-aafc-61fba4916f2d
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 487fb798-07bd-4f35-aafc-61fba4916f2d
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # 487fb798-07bd-4f35-aafc-61fba4916f2d
 
 This model is a fine-tuned version of [unsloth/tinyllama-chat](https://huggingface.co/unsloth/tinyllama-chat) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.1998
 
 ## 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        | 0.0009 | 1    | 1.4073          |
+| 1.3208        | 0.0086 | 10   | 1.3742          |
+| 1.3314        | 0.0172 | 20   | 1.2659          |
+| 1.218         | 0.0258 | 30   | 1.2177          |
+| 1.2501        | 0.0344 | 40   | 1.2021          |
+| 1.2207        | 0.0430 | 50   | 1.1998          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f2884f6c71535cec70ae30eabf080840f92bbae584d99dc6e0b455d39f42e7fb
+oid sha256:0c83bfb25e41c920e475b7f036bdcf75dd922285e388bfce1a14cadb80245269
 size 25342042