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/323b4dfdc5209b50_train_data.json
 model_type: AutoModelForCausalLM
@@ -93,7 +93,7 @@ wandb_name: 024a5686-f4d5-4c43-b234-6a0bea4dde4d
 wandb_project: Gradients-On-Demand
 wandb_run: your_name
 wandb_runid: 024a5686-f4d5-4c43-b234-6a0bea4dde4d
-warmup_steps: 1
+warmup_steps: 10
 weight_decay: 0.0
 xformers_attention: null
 
@@ -104,6 +104,8 @@ xformers_attention: null
 # 024a5686-f4d5-4c43-b234-6a0bea4dde4d
 
 This model is a fine-tuned version of [JackFram/llama-68m](https://huggingface.co/JackFram/llama-68m) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: nan
 
 ## 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        | 0.0001 | 1    | nan             |
+| 0.0           | 0.0015 | 10   | nan             |
+| 1.8638        | 0.0029 | 20   | nan             |
+| 0.0           | 0.0044 | 30   | nan             |
+| 0.0           | 0.0058 | 40   | nan             |
+| 0.0           | 0.0073 | 50   | nan             |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1c1a3b8db92daeab859dc71a8dc0d0d9162671c50d7979210861996de3bb4786
+oid sha256:ff8d38bf4768a78f1624cfae3f8a87fbef1450e1866437dacf9d5d01abea3e97
 size 1140674