3sara/colpali_italian_documents

README.md

@@ -1,5 +1,5 @@
 ---
-library_name: transformers
+library_name: peft
 license: gemma
 base_model: vidore/colpaligemma-3b-pt-448-base
 tags:
@@ -43,7 +43,7 @@ The following hyperparameters were used during training:
 - optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
 - lr_scheduler_type: linear
 - lr_scheduler_warmup_steps: 100
-- num_epochs: 2
+- num_epochs: 4
 
 ### Training results
 
@@ -54,7 +54,8 @@ The following hyperparameters were used during training:
 
 ### Framework versions
 
+- PEFT 0.15.2
 - Transformers 4.51.3
 - Pytorch 2.7.1+cu126
 - Datasets 3.6.0
-- Tokenizers 0.21.1
+- Tokenizers 0.21.1

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a2bfc8902a69c219ecc30cedb75eced5ea8c283399e77d90ac8b0653f6e571c5
-size 157071680
+oid sha256:dc561df9b5357edf39b1b59e1fe8bae061a0981dec12cbb389886b3eea62423a
+size 157210936

trainer_state.json

@@ -2,9 +2,9 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.9846153846153847,
+  "epoch": 3.994871794871795,
   "eval_steps": 500,
-  "global_step": 194,
+  "global_step": 388,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -81,19 +81,89 @@
       "step": 180
     },
     {
-      "epoch": 1.9846153846153847,
-      "step": 194,
-      "total_flos": 4.66227954266256e+16,
-      "train_loss": 0.15247911091932317,
-      "train_runtime": 4721.5867,
-      "train_samples_per_second": 0.66,
-      "train_steps_per_second": 0.041
+      "epoch": 2.0615384615384613,
+      "grad_norm": 0.1917697638273239,
+      "learning_rate": 2.4083769633507854e-05,
+      "loss": 0.0433,
+      "step": 200
+    },
+    {
+      "epoch": 2.2666666666666666,
+      "grad_norm": 0.0892116129398346,
+      "learning_rate": 2.934027777777778e-05,
+      "loss": 0.0348,
+      "step": 220
+    },
+    {
+      "epoch": 2.471794871794872,
+      "grad_norm": 1.184501051902771,
+      "learning_rate": 2.5868055555555558e-05,
+      "loss": 0.0639,
+      "step": 240
+    },
+    {
+      "epoch": 2.676923076923077,
+      "grad_norm": 0.4964236617088318,
+      "learning_rate": 2.2395833333333337e-05,
+      "loss": 0.1196,
+      "step": 260
+    },
+    {
+      "epoch": 2.882051282051282,
+      "grad_norm": 8.024321556091309,
+      "learning_rate": 1.8923611111111112e-05,
+      "loss": 0.0912,
+      "step": 280
+    },
+    {
+      "epoch": 3.0923076923076924,
+      "grad_norm": 0.8915512561798096,
+      "learning_rate": 1.545138888888889e-05,
+      "loss": 0.0188,
+      "step": 300
+    },
+    {
+      "epoch": 3.2974358974358973,
+      "grad_norm": 0.3651086091995239,
+      "learning_rate": 1.1979166666666667e-05,
+      "loss": 0.0696,
+      "step": 320
+    },
+    {
+      "epoch": 3.5025641025641026,
+      "grad_norm": 7.531687259674072,
+      "learning_rate": 8.506944444444445e-06,
+      "loss": 0.0439,
+      "step": 340
+    },
+    {
+      "epoch": 3.707692307692308,
+      "grad_norm": 16.617700576782227,
+      "learning_rate": 5.034722222222222e-06,
+      "loss": 0.0434,
+      "step": 360
+    },
+    {
+      "epoch": 3.9128205128205127,
+      "grad_norm": 0.07135667651891708,
+      "learning_rate": 1.5625e-06,
+      "loss": 0.0137,
+      "step": 380
+    },
+    {
+      "epoch": 3.994871794871795,
+      "step": 388,
+      "total_flos": 9.34866912820896e+16,
+      "train_loss": 0.0268674089582925,
+      "train_runtime": 4397.2831,
+      "train_samples_per_second": 1.417,
+      "train_steps_per_second": 0.088
     }
   ],
   "logging_steps": 20,
-  "max_steps": 194,
+  "max_steps": 388,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 2,
+  "num_train_epochs": 4,
   "save_steps": 200,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -107,7 +177,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 4.66227954266256e+16,
+  "total_flos": 9.34866912820896e+16,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:227a5b36405ddebedf1aae06096df20e366d504297b28bc40e00a0abd0e7abf1
+oid sha256:0ef95505f4cf0ce88ff57f2dd27802631727fca8c4b58453a7f10d2747fd6259
 size 5713