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/fc8ff42f89cfb652_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 5fb6cf04-0c51-4a70-86ab-e22dc06c9d7a
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 5fb6cf04-0c51-4a70-86ab-e22dc06c9d7a
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # c08aa417-614f-3d99-f6ee-22d35c3bf7cf
 
 This model is a fine-tuned version of [EleutherAI/pythia-160m](https://huggingface.co/EleutherAI/pythia-160m) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 3.0722
 
 ## Model description
 
@@ -130,14 +132,19 @@ 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        | 0.0002 | 1    | 3.2321          |
+| 13.0763       | 0.0016 | 10   | 3.2158          |
+| 13.7691       | 0.0033 | 20   | 3.1425          |
+| 12.5609       | 0.0049 | 30   | 3.0999          |
+| 12.7656       | 0.0065 | 40   | 3.0797          |
+| 11.9955       | 0.0081 | 50   | 3.0722          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d2185c3761771dcb7dbd60cd161adc908e6fecf11c29b3fdc1fae4c25957af7b
+oid sha256:aec3f368e33306e3d1e663394d80dc0cb9ff5573cfed1cc81f50c02165ab1169
 size 4753742