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/fb0d93ffd295c2a8_train_data.json
 model_type: AutoModelForCausalLM
@@ -89,7 +89,7 @@ wandb_name: 7a5a77d7-23c7-4fa5-91b5-25fb954aebc0
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 7a5a77d7-23c7-4fa5-91b5-25fb954aebc0
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -100,6 +100,8 @@ xformers_attention: null
 # 6a675f5b-c4e9-4aa2-ea2c-5d906bf3bf4e
 
 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: 2.1013
 
 ## 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        | 0.0092 | 1    | 2.4488          |
+| 8.1093        | 0.0924 | 10   | 2.2831          |
+| 8.3922        | 0.1848 | 20   | 2.1757          |
+| 8.6908        | 0.2771 | 30   | 2.1262          |
+| 7.8807        | 0.3695 | 40   | 2.1058          |
+| 7.9822        | 0.4619 | 50   | 2.1013          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fa08b98932470795d3e6761b30f931481ab862540311f3f3f00c7eaf86df01cc
+oid sha256:06a0202c0ecb6addbf1ab5f8585777081fbf5e071e51eec94858bb3e24576fcb
 size 84047370