End of training

README.md

@@ -64,7 +64,7 @@ lora_model_dir: null
 lora_r: 8
 lora_target_linear: true
 lr_scheduler: cosine
-max_steps: 17
+max_steps: 15
 micro_batch_size: 2
 mlflow_experiment_name: /tmp/559d5227401ea00d_train_data.json
 model_type: AutoModelForCausalLM
@@ -103,7 +103,7 @@ xformers_attention: null
 
 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: 10.3726
+- Loss: 10.3728
 
 ## Model description
 
@@ -131,16 +131,16 @@ The following hyperparameters were used during training:
 - optimizer: Use adamw_bnb_8bit 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: 10
-- training_steps: 17
+- training_steps: 15
 
 ### Training results
 
 | Training Loss | Epoch  | Step | Validation Loss |
 |:-------------:|:------:|:----:|:---------------:|
 | 10.3682       | 0.0005 | 1    | 10.3733         |
-| 10.3736       | 0.0024 | 5    | 10.3732         |
-| 10.3688       | 0.0047 | 10   | 10.3729         |
-| 10.371        | 0.0071 | 15   | 10.3726         |
+| 10.3764       | 0.0019 | 4    | 10.3733         |
+| 10.3747       | 0.0038 | 8    | 10.3731         |
+| 10.3756       | 0.0056 | 12   | 10.3728         |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e60655c822606e697981536e35386c2ccb4e60a7f873c37aafe3f0d7d84f4345
+oid sha256:13c0e5ae132ffa554f3c57c0bb84c9358f978fd6f2c48df8b5eed70ac911886e
 size 33666