End of training

README.md

@@ -58,7 +58,7 @@ 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/274f8441926ca85a_train_data.json
 model_type: AutoModelForCausalLM
@@ -85,7 +85,7 @@ wandb_name: f80e2174-ad87-429f-9adc-940eb3e5d75f
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: f80e2174-ad87-429f-9adc-940eb3e5d75f
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -96,6 +96,8 @@ xformers_attention: null
 # f80e2174-ad87-429f-9adc-940eb3e5d75f
 
 This model is a fine-tuned version of [fxmarty/tiny-llama-fast-tokenizer](https://huggingface.co/fxmarty/tiny-llama-fast-tokenizer) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 10.3731
 
 ## Model description
 
@@ -122,14 +124,19 @@ 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        | 0.0001 | 1    | 10.3742         |
+| 10.3743       | 0.0015 | 10   | 10.3740         |
+| 10.3757       | 0.0030 | 20   | 10.3735         |
+| 10.3744       | 0.0045 | 30   | 10.3732         |
+| 10.3724       | 0.0060 | 40   | 10.3731         |
+| 10.3729       | 0.0075 | 50   | 10.3731         |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f7f3f02a1cf3b3d7df3af12d3da9a7cfc33d33f83a015a40e2f50f8cf302821e
+oid sha256:8f541dffdfce63fbf1716dc0aa374b153d44a01ef6395efa59136a08fa75923b
 size 33666