End of training

README.md

@@ -3,9 +3,10 @@ library_name: peft
 license: apache-2.0
 base_model: Artples/L-MChat-7b
 tags:
+- axolotl
 - generated_from_trainer
 model-index:
-- name: miner_id_24
+- name: f356bf41-5032-425b-8da4-284285a6d4b3
   results: []
 ---
 
@@ -40,10 +41,10 @@ datasets:
     system_prompt: ''
 debug: null
 deepspeed: null
-early_stopping_patience: 
+early_stopping_patience: 2
 early_stopping_threshold: 1.0e-05
 eval_max_new_tokens: 128
-eval_steps: 
+eval_steps: 138
 eval_strategy: null
 flash_attention: true
 fp16: null
@@ -52,7 +53,7 @@ fsdp_config: null
 gradient_accumulation_steps: 4
 gradient_checkpointing: false
 group_by_length: false
-hub_model_id: 
+hub_model_id: mrferr3t/f356bf41-5032-425b-8da4-284285a6d4b3
 hub_repo: null
 hub_strategy: checkpoint
 hub_token: null
@@ -60,7 +61,7 @@ learning_rate: 0.0004
 load_in_4bit: false
 load_in_8bit: false
 local_rank: null
-logging_steps: 
+logging_steps: 138
 lora_alpha: 16
 lora_dropout: 0.05
 lora_fan_in_fan_out: null
@@ -68,7 +69,7 @@ lora_model_dir: null
 lora_r: 8
 lora_target_linear: true
 lr_scheduler: cosine
-max_steps: 50
+max_steps: null
 micro_batch_size: 9
 mlflow_experiment_name: /tmp/159279c5560d1ca0_train_data.json
 model_type: AutoModelForCausalLM
@@ -79,7 +80,7 @@ pad_to_sequence_len: true
 resume_from_checkpoint: null
 s2_attention: null
 sample_packing: false
-save_steps: 
+save_steps: 138
 saves_per_epoch: 0
 sequence_len: 512
 special_tokens:
@@ -89,7 +90,7 @@ tf32: false
 tokenizer_type: AutoTokenizer
 train_on_inputs: false
 trust_remote_code: true
-val_set_size: 0.0
+val_set_size: .05000000
 wandb_entity: null
 wandb_mode: disabled
 wandb_name: 0da85ac8-626a-4340-8d64-a50f9017e723
@@ -104,9 +105,11 @@ xformers_attention: null
 
 </details><br>
 
-# miner_id_24
+# f356bf41-5032-425b-8da4-284285a6d4b3
 
 This model is a fine-tuned version of [Artples/L-MChat-7b](https://huggingface.co/Artples/L-MChat-7b) on the None dataset.
+It achieves the following results on the evaluation set:
+- Loss: 0.8026
 
 ## Model description
 
@@ -134,10 +137,16 @@ 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: 100
-- training_steps: 50
+- num_epochs: 100
 
 ### Training results
 
+| Training Loss | Epoch  | Step | Validation Loss |
+|:-------------:|:------:|:----:|:---------------:|
+| No log        | 0.0015 | 1    | 1.1178          |
+| 3.1374        | 0.2108 | 138  | 0.7750          |
+| 3.0967        | 0.4215 | 276  | 0.7873          |
+| 3.1177        | 0.6323 | 414  | 0.8026          |
 
 
 ### Framework versions

adapter_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e2d1bfd31bcbb7c6ede8105802f3019ef9d52d6a57b5b200b45f6b1d6baa82ba
+oid sha256:9990e0e744a6b31e416b1cb206f26ebc6d1304f61fa4266e1ac7e6a9a9a9ee0f
 size 84047370

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:59c37ad0118d4ba73506a07d36cc16d598775c6f4156fcf098513e676eae46d7
+oid sha256:cc34008e5a29a99fc60ab96d896178d2b4f20a21f213d20b84d7205e7d1e51f4
 size 83945296