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: 8
 mlflow_experiment_name: /tmp/2d8416ab23c11ed2_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 8d71750f-4381-4439-b4c9-b191859e6304
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 8d71750f-4381-4439-b4c9-b191859e6304
-warmup_steps: 1
+warmup_steps: 2
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # 8d71750f-4381-4439-b4c9-b191859e6304
 
 This model is a fine-tuned version of [Qwen/Qwen2.5-7B](https://huggingface.co/Qwen/Qwen2.5-7B) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.4498
 
 ## Model description
 
@@ -129,13 +131,18 @@ The following hyperparameters were used during training:
 - 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
+- training_steps: 45
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | No log        | 0.0222 | 1    | 1.8411          |
+| No log        | 0.2    | 9    | 1.0835          |
+| 1.536         | 0.4    | 18   | 0.5607          |
+| 0.689         | 0.6    | 27   | 0.4736          |
+| 0.4522        | 0.8    | 36   | 0.4535          |
+| 0.443         | 1.0    | 45   | 0.4498          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:18eb32d01302f279ee978f7d4d08853a5c695b3cc99a48eb601a9f919ed29a3b
+oid sha256:bb36bafd4851f9994cdef11c1ffd22ea5048801078fda3598e684198eaf29076
 size 80881450