End of training

README.md

@@ -65,7 +65,7 @@ lora_model_dir: null
 lora_r: 8
 lora_target_linear: true
 lr_scheduler: cosine
-max_steps: 50
+max_steps: 1
 micro_batch_size: 8
 mlflow_experiment_name: /tmp/33fea428c3a9af78_train_data.json
 model_type: AutoModelForCausalLM
@@ -90,7 +90,7 @@ wandb_name: 16c5210f-4d0c-41b2-ac55-4e66607816eb
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 16c5210f-4d0c-41b2-ac55-4e66607816eb
-warmup_steps: 2
+warmup_steps: 1
 weight_decay: 0.0
 xformers_attention: null
 
@@ -101,8 +101,6 @@ xformers_attention: null
 # 16c5210f-4d0c-41b2-ac55-4e66607816eb
 
 This model is a fine-tuned version of [jhflow/mistral7b-lora-multi-turn-v2](https://huggingface.co/jhflow/mistral7b-lora-multi-turn-v2) on the None dataset.
-It achieves the following results on the evaluation set:
-- Loss: 0.5208
 
 ## Model description
 
@@ -130,18 +128,13 @@ The following hyperparameters were used during training:
 - 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        | 0.0158 | 1    | 2.9574          |
-| 25.785        | 0.1583 | 10   | 0.5463          |
-| 8.6208        | 0.3165 | 20   | 0.5284          |
-| 8.4627        | 0.4748 | 30   | 0.5255          |
-| 8.3993        | 0.6330 | 40   | 0.5219          |
-| 8.3532        | 0.7913 | 50   | 0.5208          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c80eeb1154825baf3b0f2e61839554136419fdffc3ee333d690f4620f62c8a46
+oid sha256:5f07266bcaee69e3c4d5f5ebf7a04ff2586fd3491538aa55e3e243aea61bd603
 size 84047370