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/ae3eb22d60dc8854_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: a0f8944d-868c-4841-8f21-f5f704129914
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: a0f8944d-868c-4841-8f21-f5f704129914
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 7daa4671-31c0-8850-6b6a-845fc664201a
 
 This model is a fine-tuned version of [EleutherAI/pythia-1b](https://huggingface.co/EleutherAI/pythia-1b) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 2.2300
 
 ## 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        | 0.0000 | 1    | 2.6682          |
+| 10.5961       | 0.0003 | 10   | 2.5611          |
+| 9.5395        | 0.0007 | 20   | 2.3522          |
+| 9.0419        | 0.0010 | 30   | 2.2562          |
+| 8.7587        | 0.0013 | 40   | 2.2344          |
+| 8.8223        | 0.0017 | 50   | 2.2300          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4ce425f2e0e858b2e8592e66503696083c6d80187c1029f1ae59cdad50b91d3e
+oid sha256:400c11ea1338f62a3008de73481f004328c299f067de9f07e2c5046e55e7c430
 size 16824138