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/eaf17fd41525fd10_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 42711116-91ed-41de-ab3c-39c3d34ba841
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 42711116-91ed-41de-ab3c-39c3d34ba841
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # 42711116-91ed-41de-ab3c-39c3d34ba841
 
 This model is a fine-tuned version of [Qwen/Qwen2.5-3B](https://huggingface.co/Qwen/Qwen2.5-3B) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.4556
 
 ## 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        | 0.0003 | 1    | 3.1154          |
+| 3.0594        | 0.0025 | 10   | 2.7583          |
+| 1.7782        | 0.0050 | 20   | 1.6193          |
+| 1.4176        | 0.0075 | 30   | 1.4966          |
+| 1.5101        | 0.0100 | 40   | 1.4598          |
+| 1.4508        | 0.0126 | 50   | 1.4556          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6f546b742dc90ab2a6c9695294841e848fa7ea436c6e621589e66bb559413e2d
+oid sha256:b556dc1dc261053a681e91971697fd683be2e21ebfd7bc617683ab127194e0b5
 size 60048618