End of training

README.md

@@ -64,7 +64,7 @@ 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/edf8d41fc746a7eb_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 279a85e6-08f9-4d65-8c15-7dd7c967c69a
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 279a85e6-08f9-4d65-8c15-7dd7c967c69a
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # 14effaf7-81aa-5e6a-ef1d-bac6e8fc4a8d
 
 This model is a fine-tuned version of [EleutherAI/pythia-14m](https://huggingface.co/EleutherAI/pythia-14m) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 5.6764
 
 ## Model description
 
@@ -128,14 +130,18 @@ 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: 42
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | No log        | 0.0241 | 1    | 5.9177          |
+| 23.9888       | 0.2169 | 9    | 5.7859          |
+| 24.0953       | 0.4337 | 18   | 5.6524          |
+| 24.1941       | 0.6506 | 27   | 5.7073          |
+| 22.7689       | 0.8675 | 36   | 5.6764          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e2e95dcc82aec0e5eec1df2b6f2914abf2ca9d05552c07c29ca6d91183926409
+oid sha256:126d3edd00e71e467f4505554df91fd70a15941ad8147fda36520ecf2774c45d
 size 410814