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/8cc0025255d845df_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 8f4a6548-1bd7-49bf-a53e-df1600f9e7f2
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 8f4a6548-1bd7-49bf-a53e-df1600f9e7f2
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 8f4a6548-1bd7-49bf-a53e-df1600f9e7f2
 
 This model is a fine-tuned version of [TinyLlama/TinyLlama_v1.1](https://huggingface.co/TinyLlama/TinyLlama_v1.1) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.2943
 
 ## 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    | 1.2212          |
+| 1.0987        | 0.0040 | 10   | 1.0264          |
+| 0.5911        | 0.0080 | 20   | 0.5056          |
+| 0.3551        | 0.0121 | 30   | 0.3383          |
+| 0.3406        | 0.0161 | 40   | 0.3012          |
+| 0.2914        | 0.0201 | 50   | 0.2943          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3f60a993387785e31ca4299890dd7f64b02847f9b02857120d660537aa3c843b
+oid sha256:95488695331e774929103fd37c22712b47803f23352df080df77e41188874a3b
 size 25342042