| Step | \n", "Training Loss | \n", "
|---|---|
| 1 | \n", "1.858400 | \n", "
| 2 | \n", "2.219100 | \n", "
| 3 | \n", "2.166500 | \n", "
| 4 | \n", "2.110400 | \n", "
| 5 | \n", "1.861500 | \n", "
| 6 | \n", "2.087000 | \n", "
| 7 | \n", "1.716900 | \n", "
| 8 | \n", "1.870300 | \n", "
| 9 | \n", "1.925700 | \n", "
| 10 | \n", "1.543900 | \n", "
| 11 | \n", "1.739100 | \n", "
| 12 | \n", "1.667500 | \n", "
| 13 | \n", "1.385400 | \n", "
| 14 | \n", "1.603800 | \n", "
| 15 | \n", "1.532600 | \n", "
| 16 | \n", "1.360100 | \n", "
| 17 | \n", "1.257300 | \n", "
| 18 | \n", "1.235200 | \n", "
| 19 | \n", "1.357300 | \n", "
| 20 | \n", "1.185600 | \n", "
| 21 | \n", "1.264300 | \n", "
| 22 | \n", "1.147700 | \n", "
| 23 | \n", "1.242400 | \n", "
| 24 | \n", "0.894500 | \n", "
| 25 | \n", "0.763000 | \n", "
| 26 | \n", "0.791100 | \n", "
| 27 | \n", "0.936900 | \n", "
| 28 | \n", "1.072800 | \n", "
| 29 | \n", "0.709500 | \n", "
| 30 | \n", "1.061000 | \n", "
| 31 | \n", "1.060600 | \n", "
| 32 | \n", "0.809000 | \n", "
| 33 | \n", "0.465900 | \n", "
| 34 | \n", "0.694900 | \n", "
| 35 | \n", "0.760500 | \n", "
| 36 | \n", "0.547000 | \n", "
| 37 | \n", "0.210900 | \n", "
| 38 | \n", "0.895100 | \n", "
| 39 | \n", "0.599700 | \n", "
| 40 | \n", "0.452900 | \n", "
| 41 | \n", "0.367400 | \n", "
| 42 | \n", "0.526600 | \n", "
| 43 | \n", "0.352700 | \n", "
| 44 | \n", "0.250300 | \n", "
| 45 | \n", "0.391100 | \n", "
| 46 | \n", "0.242700 | \n", "
| 47 | \n", "0.340000 | \n", "
| 48 | \n", "0.376300 | \n", "
| 49 | \n", "0.290100 | \n", "
| 50 | \n", "0.231300 | \n", "
| 51 | \n", "0.713900 | \n", "
| 52 | \n", "0.345700 | \n", "
| 53 | \n", "0.316500 | \n", "
| 54 | \n", "0.715300 | \n", "
| 55 | \n", "0.670600 | \n", "
| 56 | \n", "0.506300 | \n", "
| 57 | \n", "0.261600 | \n", "
| 58 | \n", "0.147300 | \n", "
| 59 | \n", "0.212300 | \n", "
| 60 | \n", "1.143500 | \n", "
" ] }, "metadata": {} } ] }, { "cell_type": "code", "source": [ "trainer_stats" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "b0uO1sWDzMur", "outputId": "ed480bda-509d-400f-ddc5-c36751717673" }, "execution_count": 40, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "TrainOutput(global_step=60, training_loss=0.9744117197891077, metrics={'train_runtime': 1206.8508, 'train_samples_per_second': 0.398, 'train_steps_per_second': 0.05, 'total_flos': 4.451323701362688e+16, 'train_loss': 0.9744117197891077, 'epoch': 0.25157232704402516})" ] }, "metadata": {}, "execution_count": 40 } ] }, { "cell_type": "code", "source": [ "test_data = pd.read_parquet(\"/content/test.parquet\")" ], "metadata": { "id": "HPhKC_02ryFZ" }, "execution_count": 27, "outputs": [] }, { "cell_type": "code", "source": [ "pergunta = test_data.iloc[816][\"perguntas\"]\n", "contexto = test_data.iloc[816][\"contexto\"]" ], "metadata": { "id": "r_J-MOuCsGhz" }, "execution_count": 28, "outputs": [] }, { "cell_type": "code", "source": [ "# alpaca_prompt = Copied from above\n", "FastLanguageModel.for_inference(model) # Enable native 2x faster inference\n", "inputs = tokenizer(\n", "[\n", " alpaca_prompt.format(\n", " contexto, # contexto\n", " pergunta, # pergunta\n", " \"\", # output - leave this blank for generation!\n", " )\n", "], return_tensors = \"pt\").to(\"cuda\")\n", "\n", "from transformers import TextStreamer\n", "text_streamer = TextStreamer(tokenizer)\n", "_ = model.generate(**inputs, streamer = text_streamer, max_new_tokens = 128)" ], "metadata": { "id": "hXWk5zLCojwF" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "model.save_pretrained(\"lora_model\") # Local saving\n", "tokenizer.save_pretrained(\"lora_model\")" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "BIgZ-eg4wckT", "outputId": "9b1df4c8-096b-4e52-e932-6c113053f382" }, "execution_count": 26, "outputs": [ { "output_type": "execute_result", "data": { "text/plain": [ "('lora_model/tokenizer_config.json',\n", " 'lora_model/special_tokens_map.json',\n", " 'lora_model/tokenizer.json')" ] }, "metadata": {}, "execution_count": 26 } ] }, { "cell_type": "markdown", "source": [ "# Testando se o modelo funciona" ], "metadata": { "id": "5Zn22E362QH1" } }, { "cell_type": "code", "source": [ "!pip install \"unsloth[colab-new] @ git+https://github.com/unslothai/unsloth.git\"\n", "!pip install --no-deps \"xformers<0.0.27\" \"trl<0.9.0\" peft accelerate bitsandbytes" ], "metadata": { "id": "Q8DWaCOC4acC" }, "execution_count": null, "outputs": [] }, { "cell_type": "code", "source": [ "from datasets import load_dataset\n", "from datasets import Dataset\n", "import pandas as pd" ], "metadata": { "id": "xXINZckd4dI1" }, "execution_count": 2, "outputs": [] }, { "cell_type": "code", "source": [ "test_data = pd.read_parquet(\"/content/test.parquet\")\n", "pergunta = test_data.iloc[816][\"perguntas\"]\n", "contexto = test_data.iloc[816][\"contexto\"]" ], "metadata": { "id": "lD5puCJx46Qm" }, "execution_count": 18, "outputs": [] }, { "cell_type": "code", "source": [ "from unsloth import FastLanguageModel\n", "import torch\n", "max_seq_length = 2048\n", "dtype = None\n", "load_in_4bit = True\n", "model_2, tokenizer_2 = FastLanguageModel.from_pretrained(\n", " model_name = \"matiusX/lamma-legis-ufam\",\n", " max_seq_length = max_seq_length,\n", " dtype = dtype,\n", " load_in_4bit = load_in_4bit,\n", ")\n", "FastLanguageModel.for_inference(model_2)" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 348, "referenced_widgets": [ "2dca3b1d91e947acb617b99ce5c87e5e", "d476ddd8731e417e967d827dc2cd3f42", "242e36b3cdf040ea889e304449ffa9f6", "797c07c1f5e441eda1ddfcd51999f16c", "e5be504837b94321a25cc23387d854e0", "45d3b5204634442aa8536300c4117959", "00ce9dcc9bfa45898e342c06a1df696b", "98f9785e737d4c5fa022fb665a06c5ae", "017d59d56fc04a5e900bbaadb819068e", "96fd1da82ca9426aaa4cceb42709be82", "2969c77faa4346b9876608cd9c18c7ac", "99baf6a9b88947a38d821def08009d47", "843b68ea575046c4a6730d25f7f84d90", "acea35586bf74dada4c6e254bb6ee2a5", "f44f67d24fb3421aace3a3b36ea035ff", "2f44172af2ad4a50ba617f06dd57d08f", "e5b5f362f57845f585788ba1dbc85682", "010d12f887c24dc2bd5c1d4f6d3c9ea1", "7771bebcd61b4eb8bca47eb47786e532", "99ebd1d5ac71495ab833dfab061d0c5e", "99cd3cedb7f84b2da044ecd6aded4033", "8daa9dce4f2840079a40e4a2cb9f0125", "9bc8fcdc34204d5991665a5e532beafb", "71e64f7759ab4c08b24c5fc1c77250e1", "d8ed6a333e4b4bfd8052be35ef532343", "f295df4ab0704ecfacbfcd02f76be328", "45e469e16a05433b96cceb99823b7c19", "3e7a773e12ab4a3fbebe6a622ce2a428", "67f1a26062bc417aa68326eefc1fbfe0", "df245c6077654be3800b8fca7903c571", "733694557ccd4d82ae452a1e87fdfcd4", "1cab96dd7036494bbb3cd2fb03ae49c1", "d0a00d226b40441ebf3ba731e342c90a", "6d233448c5604b5b919a36e420c778fd", "5b05b480a5284e87bbb8cf86c6123e8f", "5e1c445f069344728c5cb96db6929e11", "2242b0573311454ebf277b9a39aabde2", "515759b9350a4313a9e231fe756f8114", "3e1eb5535127407ab9c339971b1dfb69", "b561dbc164ec4815a1d1b7cf7b150d9b", "bdcb0cd08d014f588cce609fc431d534", "cb1ea1cf04f84ea486c95a3379c55823", "e654de62dc5e42408ba1f410d305f353", "fa16522af6b44e118bc802a099d69137", "d85f24809a074a44a32493c2341056fe", "9afd2e34c9534accb6946973852975d9", "73b2024cf0b54532a26c523b91cffb64", "5673d08edac64973b06c93022e58dc19", "526de9a2d69c4a498dd3835e5299cf3e", "867960ba4f014a13a8bd813abd80b66a", "5cfdde0596334625a38c46083d360a36", "c50d39dee8504a13b4082f3c9f0a733a", "f68bd6bd4bec4c53bd92e7621999b9eb", "3c96bade2a8a45c3a9dc814d932fba10", "9a7c1c86dde04a7282a3ab433f62ec34", "030f0cafdd2f43f49196baec7b63cf39", "9ae337945ea244cca1a1bd450223d8a0", "8a7ce3fc2c804c26b6c0d8b3b366d302", "ea22b2b2151e45faa0d214c6e92ac0ed", "f54d119f51604846916d9e4c33c5295a", "f4ecc536cc83497d9b78cec8f5e26c09", "ae1943703891446096cc189b2b85820c", "56e4a64c393e454d80a06c503a3513ae", "097f445dea974d4583cd408e4666f0a4", "970a3c2176384e06a87403b3a0660fdb", "ada4af9711ed4639acc7af009f3c99f5" ] }, "id": "0VSIJPZm2R7t", "outputId": "f58ea889-e009-4d0c-c293-37caaf7c27c1" }, "execution_count": 3, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "🦥 Unsloth: Will patch your computer to enable 2x faster free finetuning.\n", "==((====))== Unsloth 2024.8: Fast Llama patching. Transformers = 4.43.4.\n", " \\\\ /| GPU: Tesla T4. Max memory: 14.748 GB. Platform = Linux.\n", "O^O/ \\_/ \\ Pytorch: 2.3.1+cu121. CUDA = 7.5. CUDA Toolkit = 12.1.\n", "\\ / Bfloat16 = FALSE. FA [Xformers = 0.0.26.post1. FA2 = False]\n", " \"-____-\" Free Apache license: http://github.com/unslothai/unsloth\n", "Unsloth: Fast downloading is enabled - ignore downloading bars which are red colored!\n" ] }, { "output_type": "display_data", "data": { "text/plain": [ "model.safetensors: 0%| | 0.00/5.70G [00:00