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/79e334ea579c1de6_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 4ee24eba-461c-473a-a3ec-61c18e08e58a
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 4ee24eba-461c-473a-a3ec-61c18e08e58a
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # d21dec84-d26b-4d7b-bdb9-048fb53f6ce4
 
 This model is a fine-tuned version of [unsloth/llama-2-7b-chat](https://huggingface.co/unsloth/llama-2-7b-chat) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 2.3058
 
 ## Model description
 
@@ -128,14 +130,19 @@ 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        | 0.0002 | 1    | 2.5546          |
+| 2.4452        | 0.0023 | 10   | 2.4671          |
+| 2.3347        | 0.0046 | 20   | 2.3614          |
+| 2.2652        | 0.0069 | 30   | 2.3215          |
+| 2.2416        | 0.0092 | 40   | 2.3079          |
+| 2.301         | 0.0115 | 50   | 2.3058          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:446f5c5ec00877f38e4846e699ecad93f8e794ec4adf4ba03f86dd5fd1af9baa
+oid sha256:ae966f476cbdb08d70117a309fca267b3c1dfd7a39af9ffcab4d36d5bccdbb7e
 size 80115210