End of training

README.md

@@ -46,7 +46,7 @@ flash_attention: true
 fp16: null
 fsdp: null
 fsdp_config: null
-gradient_accumulation_steps: 16
+gradient_accumulation_steps: 4
 gradient_checkpointing: false
 group_by_length: false
 hub_model_id: tuanna08go/5c0d2573-80e9-431f-aec9-7c4bd200fb15
@@ -57,7 +57,7 @@ learning_rate: 0.0001
 load_in_4bit: false
 load_in_8bit: false
 local_rank: null
-logging_steps: 10
+logging_steps: 5
 lora_alpha: 16
 lora_dropout: 0.05
 lora_fan_in_fan_out: null
@@ -65,8 +65,8 @@ lora_model_dir: null
 lora_r: 8
 lora_target_linear: true
 lr_scheduler: cosine
-max_steps: 50
-micro_batch_size: 8
+max_steps: 1
+micro_batch_size: 2
 mlflow_experiment_name: /tmp/b460bbb94c82e6c0_train_data.json
 model_type: AutoModelForCausalLM
 num_epochs: 1
@@ -90,7 +90,7 @@ wandb_name: 5c0d2573-80e9-431f-aec9-7c4bd200fb15
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 5c0d2573-80e9-431f-aec9-7c4bd200fb15
-warmup_steps: 2
+warmup_steps: 1
 weight_decay: 0.0
 xformers_attention: null
 
@@ -101,8 +101,6 @@ xformers_attention: null
 # 5c0d2573-80e9-431f-aec9-7c4bd200fb15
 
 This model is a fine-tuned version of [unsloth/Mistral-Nemo-Instruct-2407](https://huggingface.co/unsloth/Mistral-Nemo-Instruct-2407) on the None dataset.
-It achieves the following results on the evaluation set:
-- Loss: 0.0633
 
 ## Model description
 
@@ -122,26 +120,21 @@ More information needed
 
 The following hyperparameters were used during training:
 - learning_rate: 0.0001
-- train_batch_size: 8
-- eval_batch_size: 8
+- train_batch_size: 2
+- eval_batch_size: 2
 - seed: 42
-- gradient_accumulation_steps: 16
-- total_train_batch_size: 128
+- gradient_accumulation_steps: 4
+- 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: 50
+- training_steps: 1
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
-| No log        | 0.0010 | 1    | 6.8952          |
-| 63.4986       | 0.0097 | 10   | 0.2036          |
-| 1.8375        | 0.0195 | 20   | 0.0665          |
-| 1.1045        | 0.0292 | 30   | 0.0668          |
-| 1.0411        | 0.0390 | 40   | 0.0667          |
-| 1.0332        | 0.0487 | 50   | 0.0633          |
+| No log        | 0.0001 | 1    | 6.9311          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:014489641ea0b243b2bb686e5591df672113a40e561fe57b9f6ecf20525e069e
+oid sha256:891c0f977f6352d116ce2e881a9b9f604037bff15a23ffc3d478184a6804190e
 size 114234314