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/62fb9d7cd79067b3_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 52ea7159-1ba4-4749-bb70-7f020821bed3
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 52ea7159-1ba4-4749-bb70-7f020821bed3
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # b5cdd4f3-b931-4757-8a65-ee27d1cdeb90
 
 This model is a fine-tuned version of [EleutherAI/pythia-70m-deduped](https://huggingface.co/EleutherAI/pythia-70m-deduped) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 131.2765
 
 ## 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        | 0.0005 | 1    | 132.6184        |
+| 556.8155      | 0.0046 | 10   | 132.5309        |
+| 532.5636      | 0.0092 | 20   | 132.0669        |
+| 590.3687      | 0.0139 | 30   | 131.4973        |
+| 545.1464      | 0.0185 | 40   | 131.3348        |
+| 510.6758      | 0.0231 | 50   | 131.2765        |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:932f6444a4e85e1409b085c0dfd887df7885747e5d70c6cc85f2d90620e1fbf9
+oid sha256:4b1a5d4e479a49b88fb7e1e34d53a01ae70da70f134f89237e6b6027382d64a6
 size 1590526