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/d8a7a0d4af9626da_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 634c0fbd-5d62-4560-86d6-2be48a10e55b
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 634c0fbd-5d62-4560-86d6-2be48a10e55b
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # 455c380e-44a0-ce9c-08fd-df3e3e5c7e84
 
 This model is a fine-tuned version of [Eurdem/Defne_llama3_2x8B](https://huggingface.co/Eurdem/Defne_llama3_2x8B) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.1360
 
 ## 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        | 0.0005 | 1    | 8.2768          |
+| 4.5357        | 0.0046 | 10   | 1.0215          |
+| 0.4614        | 0.0091 | 20   | 0.2140          |
+| 0.1079        | 0.0137 | 30   | 0.1602          |
+| 0.1248        | 0.0182 | 40   | 0.1483          |
+| 0.2781        | 0.0228 | 50   | 0.1360          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:92085071e4e7f04f0feeccd5bdbc2de22836e0c88ec5946052f951ad778fd62c
+oid sha256:546f3464b3c73703e636a108fe10192f51f8f1404a7bc914ca1674ab09263e4d
 size 144965450