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/e9a5de46d030ae07_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 01f4268a-9c46-4354-87d8-b3828851bd8b
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 01f4268a-9c46-4354-87d8-b3828851bd8b
-warmup_steps: 1
+warmup_steps: 2
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 01f4268a-9c46-4354-87d8-b3828851bd8b
 
 This model is a fine-tuned version of [JackFram/llama-160m](https://huggingface.co/JackFram/llama-160m) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 2.8006
 
 ## Model description
 
@@ -131,13 +133,18 @@ 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: 50
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | No log        | 0.0008 | 1    | 3.1399          |
+| 2.8562        | 0.0076 | 10   | 3.0681          |
+| 2.7322        | 0.0152 | 20   | 2.9421          |
+| 2.6651        | 0.0228 | 30   | 2.8514          |
+| 2.5504        | 0.0304 | 40   | 2.8079          |
+| 2.5387        | 0.0380 | 50   | 2.8006          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d714eb961f76e6de74901bd4d680c073c69b9bb65f9ba47116def2fc6a777927
+oid sha256:902e0c61216d45e248d723cac2db8460b6e130150cc1b63bd575d626b4759463
 size 6843306