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/bf638221d9621524_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 53ae6792-8a1c-4aad-b0e9-56452c4cfa48
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 53ae6792-8a1c-4aad-b0e9-56452c4cfa48
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # 53ae6792-8a1c-4aad-b0e9-56452c4cfa48
 
 This model is a fine-tuned version of [bigscience/bloomz-560m](https://huggingface.co/bigscience/bloomz-560m) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.8423
 
 ## Model description
 
@@ -128,14 +130,19 @@ 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        | 0.0006 | 1    | 2.1830          |
+| 9.0873        | 0.0056 | 10   | 2.0437          |
+| 7.622         | 0.0113 | 20   | 1.9020          |
+| 7.6614        | 0.0169 | 30   | 1.8544          |
+| 8.0502        | 0.0226 | 40   | 1.8465          |
+| 7.3733        | 0.0282 | 50   | 1.8423          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b34d7a7625edec8e939d6314a89abaf4c26966250bfbf751ed3c0e612f541ff6
+oid sha256:22473294757a0facc5693cbac6b9b84e0543e8170a76f2e1e4a2622465b4567d
 size 12653386