"
]
},
"metadata": {},
"execution_count": 10
}
]
},
{
"cell_type": "markdown",
"source": [
"## STEP 2: LOAD AND PRE-PROCESS DRUG REVIEWS DATASET"
],
"metadata": {
"id": "LGYKu-eF4ZvJ"
}
},
{
"cell_type": "code",
"source": [
"dataset_name = \"Zakia/drugscom_reviews\""
],
"metadata": {
"id": "v_FNuZmvtsY1"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"train_ds = load_dataset(dataset_name, split=\"train\")"
],
"metadata": {
"id": "DyK71VIut7kf",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 241,
"referenced_widgets": [
"159a87d01d374fe2878d7ab3c6844d93",
"0ed37c139fef4940aed6982a58aeb792",
"69ff48a64d57439cbd35b33efb50598a",
"b9fbc3ccaf964faabceb5e3d19bfa1c5",
"c20bac98a15b413484747e57a4715d09",
"9036d81ba2b643e3b4617e309aa2cdb7",
"eb498ea550d340ec9ea76501b215964a",
"fdab1d025be743b39608d110d08320af",
"8ff5e445704b400e9796602035701efb",
"636e5430653548d3b5b074d5e1f4a995",
"741606d62dc84c3d84114e2f9d5195df",
"07f8ae49ba55421f9c7efb43a8ddcfba",
"e3621468fa00474db14b7f04e108ea2b",
"fbf6c587e08e480192fdaca81c6255d7",
"7e8e8276be4148d488a129c21b8d1deb",
"d0cd6b570c654108b35ab4dbe3a1afca",
"ab624303b997454dab280ffbf6b65fc6",
"f88df2713a514ef58c9f2a1b867aeb57",
"cd00c6c6d3a04515a8696a6429446568",
"a50fb77e114945b587c1e07eeec989b9",
"a27d06511e114ff0827d4e34cf196526",
"3a2fe9cda338440181aec790f0b782d0",
"b3efcb2bea9744fc83a3129470cf339c",
"803b7a7993f04b389f1081317ec5b90d",
"65e78c4f085048a592ab6e183a45618a",
"6e2f337994444ca6a2744a29eeeae7bf",
"4d173a09af2c4c1fa5080b1409748b8e",
"9c60750d0ec1432ba5579ed8aafa5847",
"ca2ca1fe04e744519f73eb5401968653",
"a76fadd308164b2b97f9d65629e0c742",
"134c1239728d46f98fca2612d8243248",
"4184a147d65d45e8a62bdc8e925012a9",
"7dd5a4b864154585a245cd099ee9c67b",
"3073f4def2ff4e6098e79c5d23de9d31",
"37b1b0fde7874fc293c2e6e3e4e80d06",
"26574f1956034fc7923f7c15332a084c",
"727eeeec187844f4986de7c3ad844095",
"60ef6a86e0c04bc99999fc3d24c8a69b",
"40aa1968a384483eb2b9330648fddf93",
"fc196f0fe6194a3895046afed86a3a0c",
"33edbb32baea41d18d261f3cd7227c35",
"25681b7edae44a16a0333ab4f430f028",
"82c04c19f6884cc3a6fd8e7d7c268ace",
"a8e1ab37b3904349af3603ef6945df29",
"f8dae3b74ea34ce2ba93581b6493f0fd",
"59837565048e4293bdd9c9e0a0b3ea3d",
"dfcc0e53e5294a96b766676f1009265a",
"fc8e19a0bf5f41449f1b8789a5ce3c6e",
"f133c720943a4487b6700f4b034b0dda",
"129b4be73ff645faaee30591458593bb",
"a73d144eae264b57aae64d3fd014aafe",
"02fb810f07d14b6491d08c4774524752",
"66d8649970374fb3b1fadb1500688135",
"a7f70dc5c7164193937efa3dfa3c4e83",
"4423c35c347149a5b13a49c6d3e97bf4",
"a1e1995378b9456ebc06cb70aaef20d0",
"9f7ce81d88b0477ebeab08369c41bf10",
"53e6a9f7290444ec941df59b6e873d2f",
"a9c373347444420da58c4df1933ca9c7",
"e09a53a1cab1427e83e80a928bf213a2",
"a08f4c99d3024e7fa4fe23bfcb72d0e1",
"469d30176679485db760119a94a7ae54",
"49b2766eaba34aa78028c862419ace35",
"b9b47942c66a4d24b4cecd4387544217",
"3188a11044284d7bbf35f6f79b43c6e7",
"4ef74319a3944683ba950e325b69b253",
"ed4dabf40a9b44e29f5048ec7b27ae9c",
"7542c74a04f74b6588b302805c1e2e95",
"ac079bdadd694b6ead90dfe147a28c60",
"039eeb984f5f4365a69139eb1ed40f74",
"60fa03127e3e41d0b87502c1453f1213",
"0c7c4513cfe54caa9d542076e1ffcba9",
"e816e9e3615246c695fff4373ebd29ee",
"706ec76aa5ac4a6a8c12a08f29176afa",
"aaf933351cbd470fba3a273c21501eec",
"b78d12c084b940d392cfa3b13e7d8ea4",
"b6f07f9ed23d4e998b1a33b4259cf06f"
]
},
"outputId": "c590b863-beed-46ba-8a1d-c9df96e50278"
},
"execution_count": null,
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
"Downloading readme: 0%| | 0.00/6.59k [00:00', '', text) # Remove HTML tags\n",
" return text"
],
"metadata": {
"id": "2GmihTdSuMPF"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"# Clean the reviews\n",
"# Apply the clean_review function in a batched manner\n",
"def clean_reviews(batch):\n",
" # Apply clean_review to each review in the batch and return the modified batch\n",
" return {\"review\": [clean_review(review) for review in batch[\"review\"]]}"
],
"metadata": {
"id": "jSEt0xnfuSQu"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"# Clean the reviews for the train dataset\n",
"train_ds = train_ds.map(clean_reviews, batched=True)"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 49,
"referenced_widgets": [
"08bce8a0de8443f8befdaa81cbd3396e",
"0ed633424ede4c8784d545d5c1cf2f45",
"b15078202f204ab497bb6713d85c9b9c",
"6cbafdfa31bf4a4ea628c07d64772ebb",
"b29767f814c447be905f7435cf14664e",
"f330447c13644bfba10cf206fb4f139a",
"3705ce91e02b4c679f8d9417d78b3342",
"5a302c72d5324f94ad1d49e2bf03b19e",
"0511f90cf66f4c7492006c98fdd55316",
"a98a853bf24e4b079b31a14e86e5214e",
"05da444268d44607ae6fe181b84f5688"
]
},
"id": "T_xn2iODIHIS",
"outputId": "de603bc7-a921-4c2a-aac5-7e3e8f673949"
},
"execution_count": null,
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
"Map: 0%| | 0/9069 [00:00 RATING_THRESHOLD # Assuming a rating above 5 is positive\n",
" is_high_usefulCount = text[\"usefulCount\"] > USEFUL_COUNT_THRESHOLD\n",
" return {\"high_quality_review\": int(is_high_rating and is_high_usefulCount)}"
],
"metadata": {
"id": "UE6saXmLuctW"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"# Apply the high_quality_review function to the train dataset\n",
"train_ds = train_ds.map(high_quality_review)"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 49,
"referenced_widgets": [
"f1b1e59a777643b9a30c7e403b90cfdf",
"20bcee7a1e134fd3a11184b20e201dea",
"931c82d855804653bdf7b09bda0f8438",
"e352778970e74b8c9424d2ea100230fa",
"59675ac2b32f4e6bab6affe670b36ade",
"80f7a63437a34be6a0e7c2c1168c3ab9",
"33415320d4d54c7086aee9a295b78652",
"b62149a22ea7477dbedb098fad927e0e",
"60d900515943462989383c892e8f372c",
"952a674afdd84870aa05630a1362892a",
"ea1ffe44863c4f279673391b1c09082a"
]
},
"id": "14aQWAKOIoFl",
"outputId": "95fc3f91-4763-4c79-f8a8-e1c0575096c4"
},
"execution_count": null,
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
"Map: 0%| | 0/9069 [00:00 512). Running this sequence through the model will result in indexing errors\n"
]
},
{
"output_type": "stream",
"name": "stdout",
"text": [
"The longest review is 672 tokens long.\n"
]
}
]
},
{
"cell_type": "code",
"source": [
"# Apply the encode function to the balanced training dataset\n",
"train_encoded = balanced_train_ds.map(encode, batched=True)"
],
"metadata": {
"id": "18v0sBsc058q",
"colab": {
"base_uri": "https://localhost:8080/",
"height": 49,
"referenced_widgets": [
"a5e3115e4bc84333a83e4b32382d382d",
"d0cd69b92863460ba237ceb8d05cc05a",
"9e1f10b34c8545e694dd5223e7991f01",
"cac87a556d1b4c44bc15635ea88a0ba1",
"e27d5e192e334e89b95d74c1fca2b2ee",
"1eaad091a2ce4f18bec0ac9bface4e3a",
"41407df8693a42978fabe25ece3b9309",
"0cf1e2bff7924380900968ed3192c2e5",
"21323711b47f4daaa06ce237c0603eca",
"ac79b4f3dc3d4448a6389d2cee919503",
"ada666273e3849179ca1292a0d36ed61"
]
},
"outputId": "69f33ea6-908f-4b59-d09a-4aae3edfa978"
},
"execution_count": null,
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
"Map: 0%| | 0/4240 [00:00"
],
"text/html": [
"\n",
" \n",
" \n",
"
\n",
" [265/265 04:15, Epoch 1/1]\n",
"
\n",
" \n",
" \n",
" | Epoch | \n",
" Training Loss | \n",
" Validation Loss | \n",
" Accuracy | \n",
"
\n",
" \n",
" \n",
" \n",
" | 1 | \n",
" 0.379900 | \n",
" 0.797809 | \n",
" 0.765751 | \n",
"
\n",
" \n",
"
"
]
},
"metadata": {}
},
{
"output_type": "execute_result",
"data": {
"text/plain": [
"TrainOutput(global_step=265, training_loss=0.18166468638294148, metrics={'train_runtime': 256.159, 'train_samples_per_second': 16.552, 'train_steps_per_second': 1.035, 'total_flos': 561661770301440.0, 'train_loss': 0.18166468638294148, 'epoch': 1.0})"
]
},
"metadata": {},
"execution_count": 65
}
]
},
{
"cell_type": "markdown",
"source": [
"## STEP 5: SAVE THE FINE-TUNED DISTILBERT MODEL: distilbert-drugscom_depression_reviews"
],
"metadata": {
"id": "EFH89vCQ0pMf"
}
},
{
"cell_type": "code",
"source": [
"from huggingface_hub import notebook_login # To log to our Hugging Face account to be able to upload models to the Hub."
],
"metadata": {
"id": "LthX-XqW5Agq"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"# Login to Hugging Face within the notebook\n",
"notebook_login()\n",
"!git config --global credential.helper store"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 145,
"referenced_widgets": [
"98f03850a01d425e82f4edd0f7218281",
"a0416f5afb62412996814f538ba5268b",
"3b2c10c8b24e4a4c88e2e4fd2564b619",
"84b1a675346d4601b73c3a6b2ebf5815",
"19c44ca8d64d47b0bef4eea7cbe73bbb",
"83bae9a984fc4f6aa689abb077a05ead",
"982b11ab3b474001b9018a4f7ab7e7ac",
"81a1021ec0944d9d9f1b33f5e9184078",
"eebe3f60174240e5a71ffc054f3bf3df",
"3ea1e55d19744eacadc64e50ccbeca99",
"79ca81748cdb42d281112c64412a6ed9",
"385c71653c9648f5b9e629ba017c9b87",
"eab6012065314537960397f508af055d",
"3afc1e733ad344e1a67d9d9673c68f72",
"528328fb552c49e48fdfeb476d61b890",
"5146404d17b8401b8afca36c896cf0ca",
"eaa051e8e6734f069bfd8779ce465373",
"798ad4546b5c4977a2f6270ae518e124",
"9ebc093be8a647a6af8fea7d04656619",
"32fd60c986dd4a97b532be2f476c33f1",
"1223c060acdf499cb571ea98332c37ac",
"b88d1d7ed93c40a58ac420b5516d2b85",
"c814dd02eb0f430b90d7a88f8f692e78",
"949735e1cfd44d50a36d29dff40fe41e",
"d9ebd1d1762d4e6ba2e7e9d601b4d2ca",
"48531ee1bb6c4ead9ab152d97be3955f",
"13d5ad5688944d05a31f7a45ac272f17",
"34b6076372d94601ab30ce849aff525f",
"d47c31a9cf674660882725a4e98f4c0e",
"56861d81a5544566a5339fbc918ad24c",
"9b7c08fdd70c48e3a957ee46fee5113d",
"2a503b9ba46e4375b1a2be180f116d93"
]
},
"id": "X_g76SB35CDE",
"outputId": "6f5198e4-afc6-4269-dcfd-7d83ffcae023"
},
"execution_count": null,
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
"VBox(children=(HTML(value='