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: 1
+max_steps: 50
 micro_batch_size: 2
 mlflow_experiment_name: /tmp/0e5911bfb0cdefe5_train_data.json
 model_type: AutoModelForCausalLM
@@ -89,7 +89,7 @@ wandb_name: c01f770d-f187-43c3-bba6-ada15f3c4f4f
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: c01f770d-f187-43c3-bba6-ada15f3c4f4f
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -100,6 +100,8 @@ xformers_attention: null
 # 5f47bb07-8836-49e1-a366-074fa0de5227
 
 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: 0.9017
 
 ## Model description
 
@@ -126,14 +128,19 @@ 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    | 1.2888          |
+| 4.7386        | 0.0025 | 10   | 1.0997          |
+| 4.1983        | 0.0049 | 20   | 0.9561          |
+| 3.8974        | 0.0074 | 30   | 0.9152          |
+| 3.1976        | 0.0098 | 40   | 0.9041          |
+| 3.305         | 0.0123 | 50   | 0.9017          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:af95dda22ef868115ac218dca50647d94d9c00f3034a7b87cdf11bd4e9c8ab8b
+oid sha256:9f01a6918774ab067eb3e52eb7ba0b0baf8337c1c08a48567ba449f70e4ebaa6
 size 84047370