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/880e71c99049087a_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: f19b24d3-3b3d-49c2-b3c5-6f9ba24fe80a
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: f19b24d3-3b3d-49c2-b3c5-6f9ba24fe80a
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 83c3a25b-326e-4237-830e-94268cc4c89a
 
 This model is a fine-tuned version of [openlm-research/open_llama_3b](https://huggingface.co/openlm-research/open_llama_3b) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.3646
 
 ## 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        | 0.0013 | 1    | 1.9066          |
+| 1.8367        | 0.0126 | 10   | 1.8138          |
+| 1.5921        | 0.0252 | 20   | 1.5588          |
+| 1.5517        | 0.0377 | 30   | 1.4221          |
+| 1.2533        | 0.0503 | 40   | 1.3717          |
+| 1.3598        | 0.0629 | 50   | 1.3646          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9eaeee909def5daf7fdae7cbb1ad059897b3ed05d9b94acb572b4df69e785e3f
+oid sha256:3b4f7107676b0a12fa4a282279b1ddf8770018afa80b0f4abb2b7473d6744864
 size 50982842