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/6a6ffe8caf871a35_train_data.json
 model_type: AutoModelForCausalLM
@@ -91,7 +91,7 @@ wandb_name: 67190ca7-1619-4b3c-be0a-d27617c861bc
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 67190ca7-1619-4b3c-be0a-d27617c861bc
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -102,6 +102,8 @@ xformers_attention: null
 # fdc0bc6c-c68b-4e38-9ca4-408b162615a1
 
 This model is a fine-tuned version of [NousResearch/Hermes-2-Pro-Mistral-7B](https://huggingface.co/NousResearch/Hermes-2-Pro-Mistral-7B) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 2.1348
 
 ## 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        | 0.0014 | 1    | 2.4857          |
+| 9.2084        | 0.0141 | 10   | 2.2296          |
+| 8.5501        | 0.0282 | 20   | 2.1698          |
+| 8.6546        | 0.0423 | 30   | 2.1445          |
+| 8.8099        | 0.0564 | 40   | 2.1368          |
+| 8.518         | 0.0705 | 50   | 2.1348          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bd4c1f1f50f4fd7b0125c88e9801240038663f6236c6376ba019e00feefecd3c
+oid sha256:0dcc72eb6d473cac3ac2dad5e6a4fbef8e16c8795b07e5983f9f9fc16fa21e6b
 size 84047370