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/d7c287339d6d7874_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: cc06f8a4-4588-4595-bf2b-d3d25c761247
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: cc06f8a4-4588-4595-bf2b-d3d25c761247
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # cc06f8a4-4588-4595-bf2b-d3d25c761247
 
 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: 0.4811
 
 ## 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        | 0.0009 | 1    | 1.7184          |
+| 1.7299        | 0.0089 | 10   | 1.4282          |
+| 0.7826        | 0.0177 | 20   | 0.7069          |
+| 0.5485        | 0.0266 | 30   | 0.5371          |
+| 0.5409        | 0.0355 | 40   | 0.4899          |
+| 0.4744        | 0.0444 | 50   | 0.4811          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f96f04a669259d69e59db8ec22af6624182e6ca655740326e637ac3b613a63a6
+oid sha256:8c2442b7384fdd3de5fe1842fd10483d4d5fbd24886b396855e3d917fad3d815
 size 50982842