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/5c780282ff890654_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 79cbc040-bab6-4d61-ad30-a7b8da55262f
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 79cbc040-bab6-4d61-ad30-a7b8da55262f
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # a9f6c785-3ddc-f8de-6df2-5eba076f81c9
 
 This model is a fine-tuned version of [tiiuae/falcon-7b](https://huggingface.co/tiiuae/falcon-7b) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 1.2944
 
 ## Model description
 
@@ -130,14 +132,19 @@ 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        | 0.0004 | 1    | 2.9609          |
+| 11.3326       | 0.0036 | 10   | 2.4845          |
+| 7.1208        | 0.0072 | 20   | 1.6261          |
+| 5.9957        | 0.0108 | 30   | 1.3870          |
+| 5.4623        | 0.0144 | 40   | 1.3095          |
+| 5.2091        | 0.0180 | 50   | 1.2944          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c6e23adeaeae396ed3531e616bfcdeef314f4c72825dc26866e9526a2326fb4
+oid sha256:3452fa53f3c0cdf0c9768d125533aaaecde5d5cc1c9744e9a087e3c921119eac
 size 65367946