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/f99bc2b131bc3ad4_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: c0192787-2382-452a-bd70-d92d11e6d747
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: c0192787-2382-452a-bd70-d92d11e6d747
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 097f0465-7545-4bd7-975a-c64f5c657c07
 
 This model is a fine-tuned version of [tiiuae/falcon-7b](https://huggingface.co/tiiuae/falcon-7b) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.4167
 
 ## 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        | 0.0004 | 1    | 1.6710          |
+| 7.0404        | 0.0044 | 10   | 1.6212          |
+| 6.3266        | 0.0089 | 20   | 1.4813          |
+| 6.0375        | 0.0133 | 30   | 1.4421          |
+| 5.7293        | 0.0177 | 40   | 1.4200          |
+| 6.0671        | 0.0221 | 50   | 1.4167          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e3df0abc56ee4a3ac387e574ad91a40e68db7a68f6845e6bf7503d8842f481e2
+oid sha256:8f2ad8c5786aab955b28fb7a942667dafa626bb238693571ebf6f131dfea26d1
 size 65367946