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: 50
+max_steps: 1
 micro_batch_size: 2
 mlflow_experiment_name: /tmp/6f3ea3c9a448686a_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: e0e7dcb0-1bcf-4185-9f32-7cd479bfc180
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: e0e7dcb0-1bcf-4185-9f32-7cd479bfc180
-warmup_steps: 10
+warmup_steps: 1
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,8 +102,6 @@ xformers_attention: null
 # e0e7dcb0-1bcf-4185-9f32-7cd479bfc180
 
 This model is a fine-tuned version of [Qwen/Qwen2.5-3B-Instruct](https://huggingface.co/Qwen/Qwen2.5-3B-Instruct) on the None dataset.
-It achieves the following results on the evaluation set:
-- Loss: 3.3513
 
 ## Model description
 
@@ -130,19 +128,14 @@ 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: 10
-- training_steps: 50
+- lr_scheduler_warmup_steps: 2
+- training_steps: 1
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | No log        | 0.0001 | 1    | 5.5551          |
-| 5.0396        | 0.0015 | 10   | 4.7971          |
-| 3.9751        | 0.0030 | 20   | 3.6436          |
-| 3.3118        | 0.0045 | 30   | 3.4259          |
-| 3.0723        | 0.0059 | 40   | 3.3575          |
-| 3.2276        | 0.0074 | 50   | 3.3513          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b5a51383c8396c9803f4f221c8366ed47bad5f1ab897a894c454fb571cbd5c42
+oid sha256:185efa5d9596d6b8ae3012006b07e0ef8d99445404c88d52447d8aeb69825ddd
 size 60048618