RestartTest

RestartTest.ipynb

@@ -0,0 +1,1502 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "aacb45c2-eecc-4ab0-983f-0f459d3eb59f",
+   "metadata": {},
+   "source": [
+    "# IDEFICS_ROCOv2 (checkpoint test)\n",
+    "\n",
+    "This notebook fine-tunes [Idefics3-8B-Llama3](https://huggingface.co/HuggingFaceM4/Idefics3-8B-Llama3) model. The source model is fine-tuned on the [Radiology Objects in Context (ROCO)](https://huggingface.co/datasets/eltorio/ROCOv2-radiology) dataset, a large-scale medical and multimodal imaging collection.  \n",
+    "\n",
+    "The fine-tuning process stores the model checkpoints on a regular basis. Re run the notebook from the last checkpoint to continue the fine-tuning process."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "d0e6780c-00e5-4617-a4e8-b76e08233dac",
+   "metadata": {
+    "executionInfo": {
+     "elapsed": 1459,
+     "status": "ok",
+     "timestamp": 1730997027344,
+     "user": {
+      "displayName": "Ronan Le Meillat",
+      "userId": "09161391957806824350"
+     },
+     "user_tz": -60
+    },
+    "id": "8F3w0kcbAMtC"
+   },
+   "outputs": [],
+   "source": [
+    "dataset_id = \"eltorio/ROCOv2-radiology\"\n",
+    "prompt= \"You are an expert radiologist certified with over 15 years of experience in diagnostic imaging, describe this image\"\n",
+    "source_model_id = \"HuggingFaceM4/Idefics3-8B-Llama3\"\n",
+    "destination_model_id = \"eltorio/IDEFICS3_ROCOv2\"\n",
+    "output_dir = \"IDEFICS3_ROCOv2\""
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "020afb19-c0ee-406b-a0ee-ba0e64aeaddd",
+   "metadata": {},
+   "source": [
+    "### Log into Hugging Face"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "cfe7c2dc-fb94-43f1-a6c8-486282886727",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Hugging Face token found in environment variable\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.\n"
+     ]
+    }
+   ],
+   "source": [
+    "from huggingface_hub import login\n",
+    "import os\n",
+    "\n",
+    "if os.environ.get('HF_TOKEN') is not None:\n",
+    "  HF_TOKEN = os.environ.get('HF_TOKEN')\n",
+    "  print(f\"Hugging Face token found in environment variable\")\n",
+    "try:\n",
+    "  import google.colab\n",
+    "  from google.colab import userdata\n",
+    "  if (userdata.get('HF_TOKEN') is not None) and (HF_TOKEN == \"\"):\n",
+    "    HF_TOKEN = userdata.get('HF_TOKEN')\n",
+    "  else:\n",
+    "    raise ValueError(\"Please set your Hugging Face token in the user data panel, or pass it as an environment variable\")\n",
+    "except ModuleNotFoundError:\n",
+    "  if HF_TOKEN is None:\n",
+    "    raise ValueError(\"Please set your Hugging Face token in the user data panel, or pass it as an environment variable\")\n",
+    "\n",
+    "login(\n",
+    "  token=HF_TOKEN,\n",
+    "  add_to_git_credential=True\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "5826da8d-e57c-434a-b856-3d22d10dd2fb",
+   "metadata": {},
+   "source": [
+    "### Load the dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "a74d47eb-f798-47cc-8d98-9fd589ee60b0",
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/",
+     "height": 1000,
+     "referenced_widgets": [
+      "9f1fb7edb868495b831e12c535c4ce77",
+      "92e501753b0c49388ce317d0cbad704c",
+      "5c0db4ce097c4a7d9235c48b71e83a76",
+      "6b491ec2f7604976b347a5feb66a2a2d",
+      "98e545ecb77b461290d24e5ccd3cc1d5",
+      "18908dfa00534f9b99fd5973c070bbdc",
+      "ee551beb730743cc8fcf98c8d467de22",
+      "38b31d54ad634e19b41544ba6fa182d6",
+      "e6600d6b8cfe429480cbdb2b993c1aa6",
+      "69d0bffcb7eb4a0d95c743a99dd06725",
+      "a784b176de684e0a8be7995f16212a3a",
+      "1bd6753e9e69481992048d1eafe5849e",
+      "3e3859d3e3c240119a0eb791e611494d",
+      "af2693b9442e4c4896396db4d8336ea5",
+      "031dd614106946d79cd0dfe67cefa023",
+      "5f2589177ecd47908828ac2b19a1d0df",
+      "58304c739def48518cd0a2e51e6c8b9f",
+      "8853d39b97ab429ba5863c0a22197886",
+      "68da45f8ef1f40e89f2ecb7ae7a2617a",
+      "ae11688c634c47d89182e70801f145c3",
+      "0ca8ec79f2dc43d0897b56a3b78ee98d",
+      "e7544f37527c4bf4b6baea5e92d03ec6",
+      "bfce5ea5eec24d8e97a6188164ed31f7",
+      "c6c6989b2b8849d29b516660c375d375",
+      "505b1265c9ea426f8bde4dabba85cc0a",
+      "e5ed0c1c6a2f4202b29a2896e8e31236",
+      "eeb11f35ceb547d9aa71fb08e6a03715",
+      "90f9e80bb8bd4c39b658420517ce6df5",
+      "45d8f2f4cd2142df90b7185ee2ed75c2",
+      "2c32cc37b949475f9231d20b3ae96680",
+      "87367e46663c4a7fbfeceb5c4ecaa0b5",
+      "d2feb51c2e8347f5912b8b1f7d6252e4",
+      "0e475163ce174be1b3d41edc3330baeb",
+      "0fa957f7e02d4494b090ee9c446d7e04",
+      "ab01d3aeb23248b0bfa3b0576e43ea70",
+      "0f23267d51f84de9b621a84cf6799bf0",
+      "52c45aa3e8e5458db2fdd04955b16864",
+      "34b464cada7144bcb3753b23e1d7765d",
+      "4388b903acbe4fd189297ce4b2fd0ff5",
+      "805a520992af4639848412359eabfff5",
+      "8ea57e014d3d40fab57c058b1328d004",
+      "944df001d5d7483a92e6bc683c09701a",
+      "8bc04b1675d14cc6baa8db4a0fe8180c",
+      "d02f2bab1f2f44d1a383a100f46b49b9",
+      "058bdcd5a23240ddb7a4efa8e8334fc5",
+      "a859d2cec2d14756a93c2a90d6e7271c",
+      "38158bc387464c18a79943cbf2ff8d82",
+      "8dd54003710c4e38915d6b99faa2627f",
+      "4e8e310530ee4a0f86ea3da6541bfea4",
+      "42499379fbf3405daa80ad24127934fc",
+      "edb31fdbbd60426ea1b95aa8aadedcba",
+      "7dbaa240dd8e4f8fa8dd010be1da1e57",
+      "1331a9fe5736466e8367f7bf9d4273d2",
+      "be86432fa7b444759028c5ad2afbb7d1",
+      "e9d0cf692c5044faa9b60ac0858d3bb4",
+      "293f5a2a539443aa99764a291405d896",
+      "2cb70998cb6749aba86f4f797366ee4c",
+      "fde431fce3714221a6fb4f25ec00fc1d",
+      "c9d337da383a4cfaae8dbf1a7bda834f",
+      "cfa6776359d94782ad0abcb0eaa544d5",
+      "2dfc1bb79179445d863c17d84acfb4c9",
+      "8c34328ed5d24f22837f73f4accb84a4",
+      "58a8969990ee4da39acfccbe6f4bab81",
+      "052493c48f8c4dddadc76e845b0eef22",
+      "04d3a35c248d46c4b40efe2d8a8adb27",
+      "4ec1ab6f09cc40638ad3028494e24a90",
+      "af495be62c0f4af89c2126c40bc29573",
+      "318d5d1dd67d4e23873a3482cb198a14",
+      "a6f0b3010b414beca703085170734b3a",
+      "dabaf608430448e594c978622e6329ea",
+      "534b998a45cb4b44975dd40a5e452013",
+      "03702c12315f4be79e32f61bae0c577a",
+      "192e6a14527349ea98ece35e3064bcfc",
+      "052904cc91134bcd9d608dd8e2e725d4",
+      "a20a026b5d374c5488100b2b8e322985",
+      "94db136a74734d4386d34d6ac1a99268",
+      "5df7c75e57dc469cab54c4f88f689b12",
+      "ee8b76aad5c9435b9cae1300bf06bc11",
+      "ede1d8b2f1fd4517b297b22b05bc7007",
+      "32e72b9ecd4f4c478ac0f5597fd80830",
+      "fcf96cc71d1341918263e61d8386c998",
+      "8e0d94ea75f6457c9e77f39cdfdad903",
+      "793dbb65d5cf4e1daf5879dd37c20dd7",
+      "7cbdad6324024877a6c319722c055708",
+      "6697cfcb147742cf9303e55fce484da1",
+      "38d7358cf4ed4d588c28bb06d2e44030",
+      "db4bdf52d8d54343a41da4cecaae8317",
+      "235f704533f14d8eb9cde8ee9d7ce347",
+      "49e1bbe36f8d46408b657315fb1378cb",
+      "7364d57f39c94338ac84e92f1c246e82",
+      "3dd1fcae8ee24ac7b54810d93be13b79",
+      "b1f982c6e36a4b0ca88f52fe298bff77",
+      "9125f8f6022c449f933a0f14d33fda0a",
+      "9c5ecd24df5d44309cd3431b15443afe",
+      "300001e07fd6428ea8e0437121d7dee0",
+      "d948e3be81d945648b23c4735cc2a39d",
+      "9558d08bb2ce45fba074623960b5e2cd",
+      "558d79f87f584c05bd00f678f3754e8a",
+      "714d7b0467b94f9f8c0fac2116636d89",
+      "1c05996cd3da4d62801b1cca484fc679",
+      "1a6e7db9705a40c592f124dc3e1bbcb5",
+      "e939831f76e948c4aabd351d9f4f5983",
+      "8ba416ec3c2c41fcb979e23dfd3d2197",
+      "95c9c17fc5a449cf934ee234be3ad456",
+      "e614d797ef214a18bbc21ac5fd5d19d1",
+      "119d54ebbf6543cc9eb46f1c5eac455d",
+      "7b9ed1105d7c411e8776ef87579b1885",
+      "5e2af1d91bb84b1ea6d9d9e78810aedd",
+      "c7350d8bc53e409a9757c07a3338a091",
+      "cdd0fd1afbf5473ea5c700a696ec2714",
+      "2711e1ecd6ba4655a65aed610b2c9bb2",
+      "1a6d5bb2bf36435cb3177426776100f1",
+      "b20003b963b341fea843d78ab5fb9536",
+      "3845e901b2254703802ffa7941100d72",
+      "d006625e56394770bfa7251e758e30cf",
+      "ced9665945454cfd9e39538db2b27d5a",
+      "f2149ee0b7454651a3d4848520b9bf14",
+      "e8e9fb9c23ab40d3bd830b2098f9dc3a",
+      "141cf7d500a74e28894a43f6d4e1460e",
+      "4586b647ccbf4aaeaa76cc18d6e665cb",
+      "7aeccd8d652e454e921c42ecb967e0f5",
+      "b49f56ae212e490987b3b36a77e5aa3b",
+      "b829d0a81fa44ea7974e5289e70c8a2d",
+      "57dda77517494ddaac9628c3f0afcfa2",
+      "6b453755737643398e2e4706a7ed7e4f",
+      "c5adb0e0fe324a9cb27ef1db3becc403",
+      "0c8b3c409bee43849ceefb2536478638",
+      "57f2c0896222444bb083b1cc2cfee55e",
+      "def963efaa98452eb42a71b180e003ef",
+      "ba000563c0b040c782ed66260bd3720e",
+      "cfaa170556f244918584103683cbcc9f",
+      "df20e5eb9c5143a388f7fb87a66ae0b1",
+      "9743ce9b636748dca9d53b8edf7bba4b",
+      "30ce39cf5ec84243b079f56a7f48c346",
+      "3cc9fddbfff6402aafa8b1f95309e868",
+      "e4127c3b1e914646a0c1c284f5dda155",
+      "1f27b696c12a413886594c6a3321132e",
+      "cf6e25dedb3342d2aa78fe790886294f",
+      "9592793d73be44ee9b94e4666b66eee8",
+      "49858f181c35493fad5a78b5714553ca",
+      "4ecc9d6bb9fe42d1822dd3c51ccd796c",
+      "d395d8bea8ec45f586687222550ffbb9",
+      "fa4182c2d5b0454899a06f01cb5a50ed",
+      "a37760b13c624d9494844be3304895ec",
+      "606988656da247d0a397f51485b2e74c",
+      "ea825d67cd2a4753af417c345fde8c27",
+      "ce8bf8faf7904743962166085d6e8cac",
+      "f5886ee593cf4eaeba77704f0f945da0",
+      "cbe25015ecef494a81efba20c8824f85",
+      "5f663c5a363741ddb98385fe6d993382",
+      "5dd10c0678844bd88c7d61628b66d884",
+      "b12122a3d354486587e2d6677c3afb93",
+      "099bf7f4825a432f8536392b9c9a92f9",
+      "f2db7c87f141450faedc81abf9c1e2b9",
+      "69a712d0f248444d8415e3187a08ea66",
+      "e84dc792c822437fa285c88a6bd5f593",
+      "4498e66146c24da9972f1be3effdae35",
+      "ea10a2d4f6714286a611dc319232be74",
+      "b1be925dd9d042c487b4ed2edc5575b0",
+      "e7736b2982ee412782de039b83d6043a",
+      "2cd745b3c6314e5eb3910db8bb060dcd",
+      "fc3f4ed6aa504b60a478d91f0e74c367",
+      "6c6dd664b1ba49bcab55033b4728a1ac",
+      "8852ca4c5ce4483fb3a68e4162fdd7c2",
+      "eda2491baaaa40a0ba478f18d919f0d2",
+      "6c2d3b5ddc464c84976ab22fd427bf64",
+      "976f3fee3e4d4bc49359e65525100cb7",
+      "234b3b4fc84b45bd825bf31edfcaa609",
+      "760917fbf98c4c64bf3ed91771f90c5b",
+      "31e7f2a17f704c0587f65ab935a6474c",
+      "b8e7ee5655274cf8a95c3546f8d35a66",
+      "f30a823e93424a9f88d60a773aa48127",
+      "8dae6c92a4874b369e709cda626b9e16",
+      "f50b6e2f01304cecb4cc3bae4b973689",
+      "b1f5135ec2ed4fb39e6670fee13ae2ac",
+      "ba6f67f2fb47472eb9306d382450a639",
+      "89ddc068529c40acb5f15e1288acbf08",
+      "159ff3152c48436099bc3935b68bc5ff",
+      "4d2cc00cbda049b493e70a34de52f31d",
+      "17fb66b43c8a4b6894e13850230dad50",
+      "f2c8bf3d270b47fc98a2a73176b38e0e",
+      "44234d57983943d5a05d594aea4e2920",
+      "75d0a848fd584e3f900930f1f0284b09",
+      "9757000fbeee463287c139ff92ec1739",
+      "674e2c8442244f47942d8a46455379f3",
+      "5d9da2ae9a444bffa51ac0eeb2466c2a",
+      "f2eca44d0ca54d47bd583c708145f2d7",
+      "602931dff9364a3bb8879726fe9ba637",
+      "9d4319a9e4534cdaab0473fdb58fd040",
+      "b10f679d7f674190a71ee70feb568f2e",
+      "4d040b9b8369407b80d9acee933ead6a",
+      "c7993393d2e24ef693f66ee024f1e149",
+      "da133d2847b94ae89220c6f18ead0020",
+      "04f2a883e2c349ac90e7242a80b27066",
+      "41f12316e4b54358bf98ddf919c897a7",
+      "54509e9406474869a2f459b3abc5aa22",
+      "0e605fc1126849a5a83f5833de8ff3ab",
+      "849bfba20fde4288ba38685cebdfb35c",
+      "048a2a63e0374bcf9208e195c421957a",
+      "b63bd174450a4b8097d2dd83306533c8",
+      "060aa178f0e44659942667115f2a0e43",
+      "13bde01a38044aa082a5f236facbeb5f",
+      "77a9761fbb764ff4b3d543afff4f26d2",
+      "a46a8c9089754927923d84ddd9ab167e",
+      "1b7190e5a2984695852e8cfa2aa5894a",
+      "fa97b3ea652d48a4b9942abb23f46259",
+      "dbeef4558d2147b289fa108ab47a2dcc",
+      "a36f2eb9970340ec96c37f2dec04d3bc",
+      "cc613172d44a42f1bb24f94660710cc3",
+      "94dee2553d274b119d0c8ef589c7f40d",
+      "04bd7ac45116475fbfcf63210e27bbc1",
+      "41884cd70c104ac38ca29e566b9e3d97",
+      "25f1f07563a040749690b9447181267e",
+      "a8da49836fcb43b990ac5f38559e6221",
+      "91e5719eb4744510a58cd1b2236afd69",
+      "709e43eb493e4808b9fb0bdf8d8c70c3",
+      "6264cbd8bcdc4cf4b98871ec6d62d7c0",
+      "ebed8d6a297a41c19bf8adb998428218",
+      "7ebc2f1ce5b84d0b91d0697849330a71",
+      "aeca0598f8e64f8bb16b5dec78f6e2cd",
+      "8a393051dda141f2a107f453cb3a21bd",
+      "b67a33887f2c457c805df04eaea25f12",
+      "501f10738ea742c59a423c00131631e2",
+      "88cb916758314187b0d880df7740ff51",
+      "957d482f0b164233a6c0884d6804e99c",
+      "cee3a6b3b1774e6e8426afe66d738678",
+      "2ec30c90fde44cfe83c8a4fec8f4f0c8",
+      "c831e1725b234647afd35bcdf5b60a48",
+      "7ce0aed330c948bcae8e8c14a5c84b1f",
+      "f8d99fe78fb2479d851bb3be2521539e",
+      "e8e91eae7b374756a873c1a011d5fecb",
+      "37052784502e46d6b50702eeb7786c5d",
+      "33f434725edc4305ab70bc12976e1ee1",
+      "7517ae4b420645e5928928cced44b1ba",
+      "8e9b3df7fb7044c5b3d53b2d05fa1237",
+      "e932dd7eaafe45288452c61ec6e69e10",
+      "ae05630c20a145ff934264a196bb53c5",
+      "cf862ffe6dd8447495d2eb24fc32c54e",
+      "8e2470267fbb4328a18071a6b1a88bdb",
+      "570f0f87519f43f08e28286e0f7cc9aa",
+      "516e1569f9c644b5b7f9680ad1836a0e",
+      "224f44f0c138439599b64b0a22e18561",
+      "2f030628be5944e59bbbc71845b459e5",
+      "6e3835d20e014fa0a156d37c7cdca45b",
+      "6562b0c2d2324f589b91f5d73b1572ea",
+      "5399443d85a941ca8c1d8de3c82d474e",
+      "caa506dfc4b840a9b12dbad5c369d2a9",
+      "98a3259b6668496191c42b4d938989b1",
+      "ce6b97944af3468da286531314b1be55",
+      "f06d40ec0b78474e8e00b8c3337b7c37",
+      "a25f7ac10ff6440ca254c9d649835282",
+      "1b213b4516b2474a81dadc8b62891b40",
+      "4e5cea218561410ab566ee3f8250c7a7",
+      "a0ca9d5fe07841c7a763dff2c09e3723",
+      "c96b087decc04e7f9570524baeee690f",
+      "634d38780680487c9fb4c5602ad49721",
+      "39614acd98c744aea09a933c28b4203e",
+      "159b2fcf518248e8b0e2e3a78c76f05a",
+      "4915320c740c4ae5baeb2d343987823e",
+      "7766e3ff19b345c782bc23c74b115fd6",
+      "4dd83966ac5340b6b5565178106a660e",
+      "a720bb7a7e154611a0fc2f4c645b86a2",
+      "67487b9f5b7e4a5b94731b21337150c1",
+      "9375eff4ffad4c5cb5a7a23868ee2ccd",
+      "8c0f22f1c7364865aa277ade8c3968bd",
+      "df76bba09bc649689cb529159a0f53ec",
+      "369ef8edc5b74e2a87a1ac280db04c76",
+      "7e4f4dd037b547a897a51d3795138ae9",
+      "797cbb6d462540b68dc7c941aebf6696",
+      "8b68d6fa26664dfba7fcb25f12eda0b2",
+      "a4d2e340bd37407cab41900028db60d1",
+      "0ec459dfb233410599d999ac0ff93cb5",
+      "9a3ea32e4533406fb3296ae024187bde",
+      "afaae56bdb7243adaef805fb12381171",
+      "1fb54fa6223d4434a87ed156bd0da59e",
+      "356dfbb75d1647ee84592bc3acdb97c4",
+      "7b98988a584a48d8835e2bafc6656244",
+      "433b1e1da2f044539d3583c4e7c87277",
+      "44f9756ad81c4cba8bbb389dfa346a21",
+      "dc92b0a050864c518bb3edc9cc95d097",
+      "d968f143ad264470b6153e285522d20e",
+      "94ba39aae77c4f5da9ed764d8bbbb787",
+      "5308870c262e4ddd9951dc773fc4ff82",
+      "afe8ec13518348588fb7c121d88bbede",
+      "ee87d5335f1a4902856d9810e7405d86",
+      "b556c220a2f644b0be33bedfc4ab37da",
+      "16f05787e29d4d9ea49607bc59c8635f",
+      "b7945dcbda23460695df21785d842e32",
+      "e25a53f84ed74e259c9c8dca384acfdd",
+      "c7c89190bbdc4ecf85b235f28cd260c4",
+      "0781b655bb6b45d895efc78d1d8800e8",
+      "2e75922cfe6e4a01ae1d754cce683b77",
+      "269f1a47aaa745d688bff8634002d6e4",
+      "21230790d0ad4aeeb06f27befaa94f57",
+      "918b680796214199ba9fda7a28b7b48a",
+      "d90021999afd4b989ac59fdedcc44df4",
+      "2c31a59376a041f9a70bdbcd95b196ac",
+      "0669c278494941728fe0795e9c6f4867",
+      "757195231ca54f5a95d0f35246c98ac1",
+      "28b98661f1094a37989eb170de5ff9fb",
+      "22b31e5d19b54aa993a6fa74a2d8676f",
+      "59382d76dd9c4a0ebeec230e0943d152",
+      "d4292fe89a44486aa120ede14d7f54f7",
+      "2b0d4caf302c4238b82ab41a8218a2ae",
+      "94850b29b8724feba55999563ca5394a",
+      "cd19571be77440b58dc378718acdcfe9",
+      "252fdfd71eb3457eb16bc2ad78bf7c92",
+      "48d07314744544eda49d7d64df228069",
+      "78c07bf46ec142e7be5efea449352368",
+      "4469befba04a46be90a1ada54be23d2f",
+      "cdb4edb6863742cbbfaee7a004f44a06",
+      "8ee664d228f24099855490ca1a6ec511",
+      "3e1cebf6d59747e69eecbebb223a46e6",
+      "8d8e96c495e1455b8bad70e4ec82e707",
+      "4b8cc06cfad64c139f680dc37ce9c0d2",
+      "0912bb3fdea04b91a18b852341586b72",
+      "bb293ee3925f4283be569177c81058cc",
+      "991c717a23104fe4a63ddbe5622bb602",
+      "42a249a3b8004fd9ac1fd6e160a76eb9",
+      "d77a6697f1094585922d186f94d4cded",
+      "d071a8c35b064d489eb3c0302389807d",
+      "8759e89091bc4ef89e6f7eec08cf0060",
+      "7c467e66ba1d4cb6a3e86732eb3808da",
+      "648baf6ccba14e5c9bd220500ac4d907",
+      "3070d7cdf5c24da982b3d067a956853d",
+      "fc543c8083454e05925f0f4a3d576a0b",
+      "f50adaf856074b539b0008a0b8eb2f8d",
+      "ec29599c1281435eb61c59b4ed36b10d",
+      "3c263202204b433ab6d2535f17c388cc",
+      "cd98582b26294c16842c6450e6745ea0",
+      "95bf9b8754eb49998481f7b7ec6de83b",
+      "93b9cd4e86034a62b03cf4a12d830561",
+      "7b3222f646c440869c4b99db56dc06f1",
+      "2f0541b6a7824b86ac2c70eae49366c3",
+      "c0ae8a54c6ef4200ab5ebbaae9d40f51",
+      "acea5f3faec04f1dbdd94b70894d675c",
+      "3602a9e748d34a3faf079bcd6bd1e4c3",
+      "1f21f3ad1e294d6f8dd6fd1b32f21f7f",
+      "f9bedd2a11ee49719830ef2132c47b4b",
+      "f1abe23757c54eb98df5e5995c02e57e",
+      "5b07414f6728479fa54ca08c9b1762d0",
+      "a908cfdd55bd4dd7bcb265b3fc543cd2",
+      "657669383cef4f1babe66f2788cf3793",
+      "99f122891e564a6187cdf1b40c630bc5",
+      "f1876ec512174551a531eedfe7f9f3f7",
+      "b43de568a9624b3dafed127ec05fb33d",
+      "a0c1768cf64d4a5e8df52aac24e3539b",
+      "1e0f5205bd884bd784cc710983cf223a",
+      "d48af41f002448198d23ed3564078478",
+      "0516ed2dcc714d21ad62450a8aefaa3c",
+      "888ca195e590415d895b60370b1dd904",
+      "f620293d10cd45afaf5ed61e2d8597ae",
+      "d918763d272549268b6f6822bc669998",
+      "2c3524bdbf2c4d3f9757dfca9f801433",
+      "81116b6fa75c4d25bd528818532227d0",
+      "cdb3c5d2068344388342c079bd45b1fe",
+      "533464f5ce984ff99ec2d53a393eb336",
+      "ae436c99493c4cc3ad4842358b3bc3ff",
+      "42fe587493cc49c28264d821f1557d35",
+      "d514b8c9d66d48039beb37e28add0a93",
+      "fb48433364dd42778a3e045f9d39c95e",
+      "0d7b3708d2894a49bfe98ebef0aa50a6",
+      "3288c16e67c64fc081a7343018a5b564",
+      "7aeb5b44ef00474284b20a8c07b356bc",
+      "a110709e1dc249d188c7e1967006bf1d",
+      "ea7f6ed74c8b40c3813cfdd1c01a5283",
+      "9deafef19f214d929337ca012dd653d0",
+      "62f32b6f17df46038b938cef9bc1b673",
+      "587a748eb72e4b9fbe1d27612fb37e79",
+      "350ab0925630433992d667a465ac1b11",
+      "9a5a69cba9b94d169bf767bb2bd75d6a",
+      "6002450696db4a15958f39f5014be380",
+      "a203f631dd374d8aa11b1face537e6be",
+      "f93d6bf133434e7899b1786e729b9f47",
+      "d5f3cb4539db47fbae84122a96acdf21",
+      "f0b3536703134a22a57b9427463b12ee",
+      "4dc1b1a7fa68433196bcb7bc30aef421",
+      "274587f016014a68bb20557d05f6e8d3",
+      "15f77c07bb6446ab9c5d3b206b1357bc",
+      "9d25562d85764b3da733616ec8bb68e1",
+      "3b13bd4ff067435586508d91868f3b63",
+      "10248cbdaa7d4ea99830d3dae0838f0f",
+      "c971e53f10b54a8faadb179df71597e7",
+      "290b76e842e84e198c70c606fe20899e",
+      "d974fa71d05148729e08d893095bad24",
+      "669ee8f1b5fc408d81c2ddd2de5f6f9c",
+      "08311eecb34641f5b602cc8a0888d86d",
+      "82c64dba67694d889f0a0c01cf3830e7",
+      "a6210ad7ceb247b0b136db6810605168",
+      "1ae53773abea471580087a823e5301c1",
+      "356430f29d3044acb0da3a7d12f611be",
+      "a14307cb17164b57bdd09e7373681272",
+      "17fe943fdb4546138856cececadbdc41",
+      "cac374bd58e14b50bac063685ed741ab",
+      "a9452c25f1664eb5a87d3e954f8768c8",
+      "17e5bc7c84f846879478f68cc26a3603"
+     ]
+    },
+    "executionInfo": {
+     "elapsed": 506620,
+     "status": "ok",
+     "timestamp": 1730997570445,
+     "user": {
+      "displayName": "Ronan Le Meillat",
+      "userId": "09161391957806824350"
+     },
+     "user_tz": -60
+    },
+    "id": "QBOsKy-jAMtF",
+    "outputId": "4ccdbf0a-f96f-42e3-8825-6d63dc4a025a"
+   },
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "0c622eb62cbe4679be84993a83e55aa6",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "README.md:   0%|          | 0.00/4.50k [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "8fcb995b9c77443eaf0c36ea1e754db6",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Resolving data files:   0%|          | 0/27 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "854ae8deb29943bd98072e19c660605c",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Resolving data files:   0%|          | 0/27 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "158297e5ad1243e99f274d2a077ef526",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Downloading data:   0%|          | 0/27 [00:00<?, ?files/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "e6eecdb2f41e425d9f2544023c3d47a8",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00000-of-00027.parquet:   0%|          | 0.00/497M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "f28ed8c569794c0ca4040a0fa9648232",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00001-of-00027.parquet:   0%|          | 0.00/504M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "4bf1572a5ba04eefac69e60b3c6e4bff",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00002-of-00027.parquet:   0%|          | 0.00/490M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "176ce0b53a6c46fda57b1bbbd1c41f60",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00003-of-00027.parquet:   0%|          | 0.00/485M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "72599b9a44a641a58967599a026d2dc0",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00004-of-00027.parquet:   0%|          | 0.00/510M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "1db1c1cf909f4c9f81f8e6be61c63c24",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00005-of-00027.parquet:   0%|          | 0.00/498M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "df31fe8468184aaa90c470d07e400681",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00006-of-00027.parquet:   0%|          | 0.00/532M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "a27d65dd71aa4180ab9417b789f048f2",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00007-of-00027.parquet:   0%|          | 0.00/482M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "a9b246fa4fd545f6bbe76923bcc9af49",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00008-of-00027.parquet:   0%|          | 0.00/497M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "0e36447bf22c4175b9ab74e006a92f8b",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00009-of-00027.parquet:   0%|          | 0.00/489M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "0c9a0e8c5873412d9a931fffa34a2d96",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00010-of-00027.parquet:   0%|          | 0.00/484M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "2a6f45d17b6b421a8fad88a71ea03460",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00011-of-00027.parquet:   0%|          | 0.00/508M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "cca490771366499cb4e65fd29e9954b8",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00012-of-00027.parquet:   0%|          | 0.00/490M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "5c46b8f7bf594f659d10cc50041873c1",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00013-of-00027.parquet:   0%|          | 0.00/499M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "e98116484e7742b1aecfd0ae36529835",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00014-of-00027.parquet:   0%|          | 0.00/499M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "11b129aa209544b78a3eb8f74a4714a2",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00015-of-00027.parquet:   0%|          | 0.00/498M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "dc898819e9a746e58b049afe1d53e990",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00016-of-00027.parquet:   0%|          | 0.00/496M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "74049ceab4974f44b4bf7175404d7fa3",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00017-of-00027.parquet:   0%|          | 0.00/498M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "ff04b945595e44bfa09949dd0aaf4130",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00018-of-00027.parquet:   0%|          | 0.00/525M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "7abcaa766ffd460da3a571e47786984a",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00019-of-00027.parquet:   0%|          | 0.00/486M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "5b2d30420a05440ba96096714f4e6cba",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00020-of-00027.parquet:   0%|          | 0.00/483M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "5e85efad7f52438ea139a2023587c4e2",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00021-of-00027.parquet:   0%|          | 0.00/495M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "fce39109516e40dea41df1aaa2b6d0c5",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00022-of-00027.parquet:   0%|          | 0.00/493M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "19a7f8b8b7644a32bb2c3fbf16fc86b9",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00023-of-00027.parquet:   0%|          | 0.00/494M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "f1ef9a20c2364aab8996143425bdf8ca",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00024-of-00027.parquet:   0%|          | 0.00/500M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "615cd7afc5bf4c669453874e4279f803",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00025-of-00027.parquet:   0%|          | 0.00/511M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "0c265d9e58a44ec0bf24ea4102fc1278",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "train-00026-of-00027.parquet:   0%|          | 0.00/517M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "4a3a2d0e1aa0438d879e200ffc86f4a0",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "validation-00000-of-00006.parquet:   0%|          | 0.00/444M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "b8c4363dc53747078e9ac666b3c043f2",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "validation-00001-of-00006.parquet:   0%|          | 0.00/424M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "b90024b300e84056b059563aeb21d6c1",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "validation-00002-of-00006.parquet:   0%|          | 0.00/428M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "dd41c92792e343ab9707622b75005f42",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "validation-00003-of-00006.parquet:   0%|          | 0.00/426M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "92d6024ae987441a926af3dddc47f55c",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "validation-00004-of-00006.parquet:   0%|          | 0.00/431M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "20ec7fe5b62144c68109a5c7c61a661e",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "validation-00005-of-00006.parquet:   0%|          | 0.00/422M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "206fcb0090a34e79906901dc66a980aa",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "test-00000-of-00006.parquet:   0%|          | 0.00/436M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "30206001bcce455d95b8715971de7101",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "test-00001-of-00006.parquet:   0%|          | 0.00/426M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "dc2ec175ed5d4b4ca447464cccc3e8b7",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "test-00002-of-00006.parquet:   0%|          | 0.00/443M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "54199780c03f4a3483ff5b0bd06c90bb",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "test-00003-of-00006.parquet:   0%|          | 0.00/432M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "58c6b85b3d8a49a389124e46fd23a928",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "test-00004-of-00006.parquet:   0%|          | 0.00/425M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "a5c2d23979ca44efa9ebd6a87f996eb8",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "test-00005-of-00006.parquet:   0%|          | 0.00/423M [00:00<?, ?B/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "0ca73023dc3443cfa33d9f03a62c21d1",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Generating train split:   0%|          | 0/59962 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "21fb216da0d0411a8ff026839d63686b",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Generating validation split:   0%|          | 0/9904 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "76e876cdd17449af908c53a8f9dd2080",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Generating test split:   0%|          | 0/9927 [00:00<?, ? examples/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "ad808a6a8a1545da9fd43f40422fe79d",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Loading dataset shards:   0%|          | 0/27 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "from datasets import load_dataset\n",
+    "\n",
+    "full_dataset = load_dataset(dataset_id,keep_in_memory=False)\n",
+    "train_dataset = full_dataset[\"train\"]\n",
+    "eval_dataset = full_dataset[\"validation\"]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "657e22f9-799c-4745-b12e-b5ff7d16a139",
+   "metadata": {},
+   "source": [
+    "### Model reloading"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "b17fdc73-6827-4b8d-8fdd-7fe47ab664ca",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "`low_cpu_mem_usage` was None, now default to True since model is quantized.\n"
+     ]
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "8bde30fde856460db776483a6ab871fd",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Loading checkpoint shards:   0%|          | 0/4 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "import torch\n",
+    "from peft import LoraConfig\n",
+    "from transformers import AutoProcessor, BitsAndBytesConfig, Idefics3ForConditionalGeneration\n",
+    "\n",
+    "DEVICE = \"cuda:0\"\n",
+    "\n",
+    "processor = AutoProcessor.from_pretrained(\n",
+    "    source_model_id,\n",
+    "    do_image_splitting=False\n",
+    ")\n",
+    "\n",
+    "model = Idefics3ForConditionalGeneration.from_pretrained(\n",
+    "    source_model_id,\n",
+    "    torch_dtype=torch.float16,\n",
+    "    quantization_config=bnb_config if USE_QLORA else None,\n",
+    ")\n",
+    "model.load_adapter(destination_model_id)\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "2c9d71e4-ab26-436b-a644-4d2c58aa46db",
+   "metadata": {
+    "id": "0JeaGZxHAMtG"
+   },
+   "source": [
+    "### Step 5: Create Data Collator for IDEFICS3 format."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "d8499a9d-a6da-4108-8325-71b0660bf422",
+   "metadata": {
+    "executionInfo": {
+     "elapsed": 426,
+     "status": "ok",
+     "timestamp": 1730998596513,
+     "user": {
+      "displayName": "Ronan Le Meillat",
+      "userId": "09161391957806824350"
+     },
+     "user_tz": -60
+    },
+    "id": "X6TWyPHaAMtH"
+   },
+   "outputs": [],
+   "source": [
+    "class MyDataCollator:\n",
+    "    def __init__(self, processor):\n",
+    "        self.processor = processor\n",
+    "        self.image_token_id = processor.tokenizer.additional_special_tokens_ids[\n",
+    "            processor.tokenizer.additional_special_tokens.index(\"<image>\")\n",
+    "        ]\n",
+    "\n",
+    "    def __call__(self, samples):\n",
+    "        texts = []\n",
+    "        images = []\n",
+    "        for sample in samples:\n",
+    "            image = sample[\"image\"]\n",
+    "            answer = sample[\"caption\"]\n",
+    "            messages = [\n",
+    "                {\n",
+    "                    \"role\": \"system\",\n",
+    "                    \"content\": [\n",
+    "                        {\"type\": \"text\", \"text\": prompt}\n",
+    "                    ]\n",
+    "\n",
+    "                },\n",
+    "                {\n",
+    "                    \"role\": \"user\",\n",
+    "                    \"content\": [\n",
+    "                        {\"type\": \"image\"},\n",
+    "                    ]\n",
+    "                },\n",
+    "                {\n",
+    "                    \"role\": \"assistant\",\n",
+    "                    \"content\": [\n",
+    "                        {\"type\": \"text\", \"text\": answer}\n",
+    "                    ]\n",
+    "                }\n",
+    "            ]\n",
+    "            text = processor.apply_chat_template(messages, add_generation_prompt=False)\n",
+    "            texts.append(text.strip())\n",
+    "            images.append([image.convert('RGB')])\n",
+    "\n",
+    "        batch = processor(text=texts, images=images, return_tensors=\"pt\", padding=True)\n",
+    "\n",
+    "        labels = batch[\"input_ids\"].clone()\n",
+    "        labels[labels == processor.tokenizer.pad_token_id] = self.image_token_id\n",
+    "        batch[\"labels\"] = labels\n",
+    "\n",
+    "        return batch\n",
+    "\n",
+    "data_collator = MyDataCollator(processor)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "e6467b63-06c1-4227-ab02-2aa5074c8231",
+   "metadata": {
+    "id": "vsq4TtIJAMtH"
+   },
+   "source": [
+    "### Step 6: Setup training parameters"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "1f06690e-4b81-4db7-882e-3f24a33350c6",
+   "metadata": {
+    "executionInfo": {
+     "elapsed": 1008,
+     "status": "ok",
+     "timestamp": 1730998601172,
+     "user": {
+      "displayName": "Ronan Le Meillat",
+      "userId": "09161391957806824350"
+     },
+     "user_tz": -60
+    },
+    "id": "Q_WKQFfoAMtH"
+   },
+   "outputs": [],
+   "source": [
+    "from transformers import TrainingArguments, Trainer\n",
+    "\n",
+    "training_args = TrainingArguments(\n",
+    "    output_dir = output_dir,\n",
+    "    overwrite_output_dir = False,\n",
+    "    auto_find_batch_size = True,\n",
+    "    learning_rate = 2e-4,\n",
+    "    fp16 = True,\n",
+    "    per_device_train_batch_size = 2,\n",
+    "    per_device_eval_batch_size = 2,\n",
+    "    gradient_accumulation_steps = 8,\n",
+    "    dataloader_pin_memory = False,\n",
+    "    save_total_limit = 3,\n",
+    "    eval_strategy = \"steps\",\n",
+    "    save_strategy = \"steps\",\n",
+    "    eval_steps = 100,\n",
+    "    save_steps = 10, # checkpoint each 10 steps\n",
+    "    resume_from_checkpoint = True,\n",
+    "    logging_steps = 5,\n",
+    "    remove_unused_columns = False,\n",
+    "    push_to_hub = False,\n",
+    "    label_names = [\"labels\"],\n",
+    "    load_best_model_at_end = False,\n",
+    "    report_to = \"none\",\n",
+    "    optim = \"paged_adamw_8bit\",\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "3dcca8d8-4d4e-49a4-9af5-4958edb9e0fc",
+   "metadata": {
+    "colab": {
+     "base_uri": "https://localhost:8080/"
+    },
+    "executionInfo": {
+     "elapsed": 426,
+     "status": "ok",
+     "timestamp": 1730998605441,
+     "user": {
+      "displayName": "Ronan Le Meillat",
+      "userId": "09161391957806824350"
+     },
+     "user_tz": -60
+    },
+    "id": "vSIo17mgAMtH",
+    "outputId": "3bebd35a-ed7f-49ee-e1bc-91594e8dcd24"
+   },
+   "outputs": [],
+   "source": [
+    "trainer = Trainer(\n",
+    "    model = model,\n",
+    "    args = training_args,\n",
+    "    data_collator = data_collator,\n",
+    "    train_dataset = train_dataset,\n",
+    "    eval_dataset = eval_dataset,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "ff256dc8-6f5a-423a-9607-81cd9ef7735f",
+   "metadata": {},
+   "outputs": [
+    {
+     "ename": "ValueError",
+     "evalue": "No valid checkpoint found in output directory (IDEFICS3_ROCOv2)",
+     "output_type": "error",
+     "traceback": [
+      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
+      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
+      "Cell \u001b[0;32mIn[14], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m trainer\u001b[38;5;241m.\u001b[39mtrain(resume_from_checkpoint\u001b[38;5;241m=\u001b[39m\u001b[38;5;28;01mTrue\u001b[39;00m)\n",
+      "File \u001b[0;32m~/.miniconda3/lib/python3.12/site-packages/transformers/trainer.py:2109\u001b[0m, in \u001b[0;36mTrainer.train\u001b[0;34m(self, resume_from_checkpoint, trial, ignore_keys_for_eval, **kwargs)\u001b[0m\n\u001b[1;32m   2107\u001b[0m     resume_from_checkpoint \u001b[38;5;241m=\u001b[39m get_last_checkpoint(args\u001b[38;5;241m.\u001b[39moutput_dir)\n\u001b[1;32m   2108\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m resume_from_checkpoint \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[0;32m-> 2109\u001b[0m         \u001b[38;5;28;01mraise\u001b[39;00m \u001b[38;5;167;01mValueError\u001b[39;00m(\u001b[38;5;124mf\u001b[39m\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mNo valid checkpoint found in output directory (\u001b[39m\u001b[38;5;132;01m{\u001b[39;00margs\u001b[38;5;241m.\u001b[39moutput_dir\u001b[38;5;132;01m}\u001b[39;00m\u001b[38;5;124m)\u001b[39m\u001b[38;5;124m\"\u001b[39m)\n\u001b[1;32m   2111\u001b[0m \u001b[38;5;28;01mif\u001b[39;00m resume_from_checkpoint \u001b[38;5;129;01mis\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28;01mNone\u001b[39;00m:\n\u001b[1;32m   2112\u001b[0m     \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m is_sagemaker_mp_enabled() \u001b[38;5;129;01mand\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mis_deepspeed_enabled \u001b[38;5;129;01mand\u001b[39;00m \u001b[38;5;129;01mnot\u001b[39;00m \u001b[38;5;28mself\u001b[39m\u001b[38;5;241m.\u001b[39mis_fsdp_enabled:\n",
+      "\u001b[0;31mValueError\u001b[0m: No valid checkpoint found in output directory (IDEFICS3_ROCOv2)"
+     ]
+    }
+   ],
+   "source": [
+    "trainer.train(resume_from_checkpoint=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "bd15f877-ed10-4a6d-8b36-ebaaab875526",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "Copy-and-paste the text below in your GitHub issue and FILL OUT the two last points.\n",
+      "\n",
+      "- `transformers` version: 4.47.0.dev0\n",
+      "- Platform: Linux-5.15.167.4-microsoft-standard-WSL2-x86_64-with-glibc2.31\n",
+      "- Python version: 3.12.7\n",
+      "- Huggingface_hub version: 0.26.2\n",
+      "- Safetensors version: 0.4.5\n",
+      "- Accelerate version: 1.1.1\n",
+      "- Accelerate config: \tnot found\n",
+      "- PyTorch version (GPU?): 2.5.1+cu124 (True)\n",
+      "- Tensorflow version (GPU?): not installed (NA)\n",
+      "- Flax version (CPU?/GPU?/TPU?): not installed (NA)\n",
+      "- Jax version: not installed\n",
+      "- JaxLib version: not installed\n",
+      "- Using distributed or parallel set-up in script?: <fill in>\n",
+      "- Using GPU in script?: <fill in>\n",
+      "- GPU type: NVIDIA GeForce RTX 2060\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "!transformers-cli env"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "f82b7c75-c859-451f-9c40-fe9a039bf769",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3 (ipykernel)",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.7"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}