o  h9@sdZddlZddlZddlZddlZddlZddZddZddZd d Z d Z d Z d dZ ddZ dBddZdZdZdZdZdZdZdZdZdZdZdZdZdZd Zd!Zeeeeeeeeeeeeeed"Zd#Zd$Z d%Z!d&Z"d'Z#d(Z$d)Z%d*Z&d+Z'd,Z(d-Z)e!e ee$e"e%e#e&e'e(e)d. Z*d/Z+d0Z,d/Z-d1Z.d2Z/d3Z0d4Z1e-e,e+e0e.e/e1d5Z2d6d7Z3ddddd8d9d:ddddddd; dd?Z5d@dAZ6dS)Dz3 Doc utilities: Utilities related to documentation NcCs@t|rdSt|}|d}t|t|}d|S)z^Return the indentation level of the start of the docstring of a class or function (or method).r)inspectisclass getsource splitlineslenlstrip)funcsource first_linefunction_def_levelr j/var/www/html/construction_image-detection-poc/venv/lib/python3.10/site-packages/transformers/utils/doc.pyget_docstring_indentation_levels   rcfdd}|S)Ncs$d|jdur |jnd|_|SN)join__doc__fndocstrr rdocstring_decorator% z1add_start_docstrings..docstring_decoratorr rrr rradd_start_docstrings$ rcr)Ncsd|jddd}d|d}t||jdur|jnd}ztdd |D}t|t|}Wn tyB}Ynw}|d kr_fd d D}t t |d }d ||}|||_|S)Nz[`.rz`]z The aa forward method, overrides the `__call__` special method. Although the recipe for forward pass needs to be defined within this function, one should call the [`Module`] instance afterwards instead of this since the former takes care of running the pre and post processing steps while the latter silently ignores them. rcss |] }|dkr|VqdS)rN)strip.0liner r r =szUadd_start_docstrings_to_model_forward..docstring_decorator..rcs"g|] }tt|dqS) )textwrapindentdedent)r!doccorrect_indentationr r Fs"zVadd_start_docstrings_to_model_forward..docstring_decorator..r$) __qualname__splitrrnextrrr StopIterationr%r&r'r)r class_nameintro current_docfirst_non_emptydoc_indentationdocs docstringrr)rr-s"     zBadd_start_docstrings_to_model_forward..docstring_decoratorr rr rr%add_start_docstrings_to_model_forward,s  r7cr)Ncs$|jdur|jndd|_|Sr)rrrrr rrQrz/add_end_docstrings..docstring_decoratorr rr rradd_end_docstringsPrr8a: Returns: [`{full_output_type}`] or `tuple(torch.FloatTensor)`: A [`{full_output_type}`] or a tuple of `torch.FloatTensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([`{config_class}`]) and inputs. a* Returns: [`{full_output_type}`] or `tuple(tf.Tensor)`: A [`{full_output_type}`] or a tuple of `tf.Tensor` (if `return_dict=False` is passed or when `config.return_dict=False`) comprising various elements depending on the configuration ([`{config_class}`]) and inputs. cCs$td|}|dur dS|dS)z.Returns the indentation in the first line of tz^(\s*)\SNrr)researchgroups)tr:r r r _get_indentjs r=cCst|}g}d}|dD](}t||kr*t|dkr$||dd|d}q ||ddd7}q ||ddtt|D]}tdd||||<td d ||||<qEd|S) z,Convert output_args_doc to display properly.r rNz^(\s+)(\S+)(\s+)z \1- **\2**\3z :\s*\n\s*(\S)z -- \1)r=r-rappendranger9subr)output_args_docr&blocks current_blockr"ir r r_convert_output_args_docps    rHc sX|j}|d}d}|t|kr.td||dur.|d7}|t|kr.td||dus|t|krDd||dd}t|}n td|jd|j d|j}|j d r^t nt }|j ||d }||} |dur| d}d}t||dkr|d7}t||dks~tt||} | |krd || fd d |D}d|} | S)zH Prepares the return part of the docstring using `output_type`. r>rz^\s*(Args|Parameters):\s*$Nz@No `Args` or `Parameters` section is found in the docstring of `zH`. Make sure it has docstring and contain either `Args` or `Parameters`.rTF)full_output_type config_classr$cs(g|]}t|dkr|n|qS)r)rr to_addr rr+s(z._prepare_output_docstrings..)rr-rr9r:rrH ValueError__name__ __module__ startswithTF_RETURN_INTRODUCTIONPT_RETURN_INTRODUCTIONformatr=) output_typerL min_indentoutput_docstringlinesrGparams_docstringrKr1resultr&r rMr_prepare_output_docstringss8        r\aJ This example uses a random model as the real ones are all very big. To get proper results, you should use {real_checkpoint} instead of {fake_checkpoint}. If you get out-of-memory when loading that checkpoint, you can try adding `device_map="auto"` in the `from_pretrained` call. a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import torch >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer( ... "HuggingFace is a company based in Paris and New York", add_special_tokens=False, return_tensors="pt" ... ) >>> with torch.no_grad(): ... logits = model(**inputs).logits >>> predicted_token_class_ids = logits.argmax(-1) >>> # Note that tokens are classified rather then input words which means that >>> # there might be more predicted token classes than words. >>> # Multiple token classes might account for the same word >>> predicted_tokens_classes = [model.config.id2label[t.item()] for t in predicted_token_class_ids[0]] >>> predicted_tokens_classes {expected_output} >>> labels = predicted_token_class_ids >>> loss = model(**inputs, labels=labels).loss >>> round(loss.item(), 2) {expected_loss} ``` a_ Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import torch >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> question, text = "Who was Jim Henson?", "Jim Henson was a nice puppet" >>> inputs = tokenizer(question, text, return_tensors="pt") >>> with torch.no_grad(): ... outputs = model(**inputs) >>> answer_start_index = outputs.start_logits.argmax() >>> answer_end_index = outputs.end_logits.argmax() >>> predict_answer_tokens = inputs.input_ids[0, answer_start_index : answer_end_index + 1] >>> tokenizer.decode(predict_answer_tokens, skip_special_tokens=True) {expected_output} >>> # target is "nice puppet" >>> target_start_index = torch.tensor([{qa_target_start_index}]) >>> target_end_index = torch.tensor([{qa_target_end_index}]) >>> outputs = model(**inputs, start_positions=target_start_index, end_positions=target_end_index) >>> loss = outputs.loss >>> round(loss.item(), 2) {expected_loss} ``` a Example of single-label classification: ```python >>> import torch >>> from transformers import AutoTokenizer, {model_class} >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> with torch.no_grad(): ... logits = model(**inputs).logits >>> predicted_class_id = logits.argmax().item() >>> model.config.id2label[predicted_class_id] {expected_output} >>> # To train a model on `num_labels` classes, you can pass `num_labels=num_labels` to `.from_pretrained(...)` >>> num_labels = len(model.config.id2label) >>> model = {model_class}.from_pretrained("{checkpoint}", num_labels=num_labels) >>> labels = torch.tensor([1]) >>> loss = model(**inputs, labels=labels).loss >>> round(loss.item(), 2) {expected_loss} ``` Example of multi-label classification: ```python >>> import torch >>> from transformers import AutoTokenizer, {model_class} >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}", problem_type="multi_label_classification") >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> with torch.no_grad(): ... logits = model(**inputs).logits >>> predicted_class_ids = torch.arange(0, logits.shape[-1])[torch.sigmoid(logits).squeeze(dim=0) > 0.5] >>> # To train a model on `num_labels` classes, you can pass `num_labels=num_labels` to `.from_pretrained(...)` >>> num_labels = len(model.config.id2label) >>> model = {model_class}.from_pretrained( ... "{checkpoint}", num_labels=num_labels, problem_type="multi_label_classification" ... ) >>> labels = torch.sum( ... torch.nn.functional.one_hot(predicted_class_ids[None, :].clone(), num_classes=num_labels), dim=1 ... ).to(torch.float) >>> loss = model(**inputs, labels=labels).loss ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import torch >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("The capital of France is {mask}.", return_tensors="pt") >>> with torch.no_grad(): ... logits = model(**inputs).logits >>> # retrieve index of {mask} >>> mask_token_index = (inputs.input_ids == tokenizer.mask_token_id)[0].nonzero(as_tuple=True)[0] >>> predicted_token_id = logits[0, mask_token_index].argmax(axis=-1) >>> tokenizer.decode(predicted_token_id) {expected_output} >>> labels = tokenizer("The capital of France is Paris.", return_tensors="pt")["input_ids"] >>> # mask labels of non-{mask} tokens >>> labels = torch.where(inputs.input_ids == tokenizer.mask_token_id, labels, -100) >>> outputs = model(**inputs, labels=labels) >>> round(outputs.loss.item(), 2) {expected_loss} ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import torch >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> outputs = model(**inputs) >>> last_hidden_states = outputs.last_hidden_state ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import torch >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> prompt = "In Italy, pizza served in formal settings, such as at a restaurant, is presented unsliced." >>> choice0 = "It is eaten with a fork and a knife." >>> choice1 = "It is eaten while held in the hand." >>> labels = torch.tensor(0).unsqueeze(0) # choice0 is correct (according to Wikipedia ;)), batch size 1 >>> encoding = tokenizer([prompt, prompt], [choice0, choice1], return_tensors="pt", padding=True) >>> outputs = model(**{{k: v.unsqueeze(0) for k, v in encoding.items()}}, labels=labels) # batch size is 1 >>> # the linear classifier still needs to be trained >>> loss = outputs.loss >>> logits = outputs.logits ``` a Example: ```python >>> import torch >>> from transformers import AutoTokenizer, {model_class} >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="pt") >>> outputs = model(**inputs, labels=inputs["input_ids"]) >>> loss = outputs.loss >>> logits = outputs.logits ``` aY Example: ```python >>> from transformers import AutoProcessor, {model_class} >>> import torch >>> from datasets import load_dataset >>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation", trust_remote_code=True) >>> dataset = dataset.sort("id") >>> sampling_rate = dataset.features["audio"].sampling_rate >>> processor = AutoProcessor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> # audio file is decoded on the fly >>> inputs = processor(dataset[0]["audio"]["array"], sampling_rate=sampling_rate, return_tensors="pt") >>> with torch.no_grad(): ... outputs = model(**inputs) >>> last_hidden_states = outputs.last_hidden_state >>> list(last_hidden_states.shape) {expected_output} ``` au Example: ```python >>> from transformers import AutoProcessor, {model_class} >>> from datasets import load_dataset >>> import torch >>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation", trust_remote_code=True) >>> dataset = dataset.sort("id") >>> sampling_rate = dataset.features["audio"].sampling_rate >>> processor = AutoProcessor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> # audio file is decoded on the fly >>> inputs = processor(dataset[0]["audio"]["array"], sampling_rate=sampling_rate, return_tensors="pt") >>> with torch.no_grad(): ... logits = model(**inputs).logits >>> predicted_ids = torch.argmax(logits, dim=-1) >>> # transcribe speech >>> transcription = processor.batch_decode(predicted_ids) >>> transcription[0] {expected_output} >>> inputs["labels"] = processor(text=dataset[0]["text"], return_tensors="pt").input_ids >>> # compute loss >>> loss = model(**inputs).loss >>> round(loss.item(), 2) {expected_loss} ``` a Example: ```python >>> from transformers import AutoFeatureExtractor, {model_class} >>> from datasets import load_dataset >>> import torch >>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation", trust_remote_code=True) >>> dataset = dataset.sort("id") >>> sampling_rate = dataset.features["audio"].sampling_rate >>> feature_extractor = AutoFeatureExtractor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> # audio file is decoded on the fly >>> inputs = feature_extractor(dataset[0]["audio"]["array"], sampling_rate=sampling_rate, return_tensors="pt") >>> with torch.no_grad(): ... logits = model(**inputs).logits >>> predicted_class_ids = torch.argmax(logits, dim=-1).item() >>> predicted_label = model.config.id2label[predicted_class_ids] >>> predicted_label {expected_output} >>> # compute loss - target_label is e.g. "down" >>> target_label = model.config.id2label[0] >>> inputs["labels"] = torch.tensor([model.config.label2id[target_label]]) >>> loss = model(**inputs).loss >>> round(loss.item(), 2) {expected_loss} ``` a Example: ```python >>> from transformers import AutoFeatureExtractor, {model_class} >>> from datasets import load_dataset >>> import torch >>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation", trust_remote_code=True) >>> dataset = dataset.sort("id") >>> sampling_rate = dataset.features["audio"].sampling_rate >>> feature_extractor = AutoFeatureExtractor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> # audio file is decoded on the fly >>> inputs = feature_extractor(dataset[0]["audio"]["array"], return_tensors="pt", sampling_rate=sampling_rate) >>> with torch.no_grad(): ... logits = model(**inputs).logits >>> probabilities = torch.sigmoid(logits[0]) >>> # labels is a one-hot array of shape (num_frames, num_speakers) >>> labels = (probabilities > 0.5).long() >>> labels[0].tolist() {expected_output} ``` a. Example: ```python >>> from transformers import AutoFeatureExtractor, {model_class} >>> from datasets import load_dataset >>> import torch >>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation", trust_remote_code=True) >>> dataset = dataset.sort("id") >>> sampling_rate = dataset.features["audio"].sampling_rate >>> feature_extractor = AutoFeatureExtractor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> # audio file is decoded on the fly >>> inputs = feature_extractor( ... [d["array"] for d in dataset[:2]["audio"]], sampling_rate=sampling_rate, return_tensors="pt", padding=True ... ) >>> with torch.no_grad(): ... embeddings = model(**inputs).embeddings >>> embeddings = torch.nn.functional.normalize(embeddings, dim=-1).cpu() >>> # the resulting embeddings can be used for cosine similarity-based retrieval >>> cosine_sim = torch.nn.CosineSimilarity(dim=-1) >>> similarity = cosine_sim(embeddings[0], embeddings[1]) >>> threshold = 0.7 # the optimal threshold is dataset-dependent >>> if similarity < threshold: ... print("Speakers are not the same!") >>> round(similarity.item(), 2) {expected_output} ``` a Example: ```python >>> from transformers import AutoImageProcessor, {model_class} >>> import torch >>> from datasets import load_dataset >>> dataset = load_dataset("huggingface/cats-image", trust_remote_code=True) >>> image = dataset["test"]["image"][0] >>> image_processor = AutoImageProcessor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = image_processor(image, return_tensors="pt") >>> with torch.no_grad(): ... outputs = model(**inputs) >>> last_hidden_states = outputs.last_hidden_state >>> list(last_hidden_states.shape) {expected_output} ``` a Example: ```python >>> from transformers import AutoImageProcessor, {model_class} >>> import torch >>> from datasets import load_dataset >>> dataset = load_dataset("huggingface/cats-image", trust_remote_code=True) >>> image = dataset["test"]["image"][0] >>> image_processor = AutoImageProcessor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = image_processor(image, return_tensors="pt") >>> with torch.no_grad(): ... logits = model(**inputs).logits >>> # model predicts one of the 1000 ImageNet classes >>> predicted_label = logits.argmax(-1).item() >>> print(model.config.id2label[predicted_label]) {expected_output} ``` )SequenceClassificationQuestionAnsweringTokenClassificationMultipleChoiceMaskedLMLMHead BaseModelSpeechBaseModelCTCAudioClassificationAudioFrameClassification AudioXVectorVisionBaseModelImageClassificationaI Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import tensorflow as tf >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer( ... "HuggingFace is a company based in Paris and New York", add_special_tokens=False, return_tensors="tf" ... ) >>> logits = model(**inputs).logits >>> predicted_token_class_ids = tf.math.argmax(logits, axis=-1) >>> # Note that tokens are classified rather then input words which means that >>> # there might be more predicted token classes than words. >>> # Multiple token classes might account for the same word >>> predicted_tokens_classes = [model.config.id2label[t] for t in predicted_token_class_ids[0].numpy().tolist()] >>> predicted_tokens_classes {expected_output} ``` ```python >>> labels = predicted_token_class_ids >>> loss = tf.math.reduce_mean(model(**inputs, labels=labels).loss) >>> round(float(loss), 2) {expected_loss} ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import tensorflow as tf >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> question, text = "Who was Jim Henson?", "Jim Henson was a nice puppet" >>> inputs = tokenizer(question, text, return_tensors="tf") >>> outputs = model(**inputs) >>> answer_start_index = int(tf.math.argmax(outputs.start_logits, axis=-1)[0]) >>> answer_end_index = int(tf.math.argmax(outputs.end_logits, axis=-1)[0]) >>> predict_answer_tokens = inputs.input_ids[0, answer_start_index : answer_end_index + 1] >>> tokenizer.decode(predict_answer_tokens) {expected_output} ``` ```python >>> # target is "nice puppet" >>> target_start_index = tf.constant([{qa_target_start_index}]) >>> target_end_index = tf.constant([{qa_target_end_index}]) >>> outputs = model(**inputs, start_positions=target_start_index, end_positions=target_end_index) >>> loss = tf.math.reduce_mean(outputs.loss) >>> round(float(loss), 2) {expected_loss} ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import tensorflow as tf >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="tf") >>> logits = model(**inputs).logits >>> predicted_class_id = int(tf.math.argmax(logits, axis=-1)[0]) >>> model.config.id2label[predicted_class_id] {expected_output} ``` ```python >>> # To train a model on `num_labels` classes, you can pass `num_labels=num_labels` to `.from_pretrained(...)` >>> num_labels = len(model.config.id2label) >>> model = {model_class}.from_pretrained("{checkpoint}", num_labels=num_labels) >>> labels = tf.constant(1) >>> loss = model(**inputs, labels=labels).loss >>> round(float(loss), 2) {expected_loss} ``` a4 Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import tensorflow as tf >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("The capital of France is {mask}.", return_tensors="tf") >>> logits = model(**inputs).logits >>> # retrieve index of {mask} >>> mask_token_index = tf.where((inputs.input_ids == tokenizer.mask_token_id)[0]) >>> selected_logits = tf.gather_nd(logits[0], indices=mask_token_index) >>> predicted_token_id = tf.math.argmax(selected_logits, axis=-1) >>> tokenizer.decode(predicted_token_id) {expected_output} ``` ```python >>> labels = tokenizer("The capital of France is Paris.", return_tensors="tf")["input_ids"] >>> # mask labels of non-{mask} tokens >>> labels = tf.where(inputs.input_ids == tokenizer.mask_token_id, labels, -100) >>> outputs = model(**inputs, labels=labels) >>> round(float(outputs.loss), 2) {expected_loss} ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import tensorflow as tf >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="tf") >>> outputs = model(inputs) >>> last_hidden_states = outputs.last_hidden_state ``` a# Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import tensorflow as tf >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> prompt = "In Italy, pizza served in formal settings, such as at a restaurant, is presented unsliced." >>> choice0 = "It is eaten with a fork and a knife." >>> choice1 = "It is eaten while held in the hand." >>> encoding = tokenizer([prompt, prompt], [choice0, choice1], return_tensors="tf", padding=True) >>> inputs = {{k: tf.expand_dims(v, 0) for k, v in encoding.items()}} >>> outputs = model(inputs) # batch size is 1 >>> # the linear classifier still needs to be trained >>> logits = outputs.logits ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> import tensorflow as tf >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="tf") >>> outputs = model(inputs) >>> logits = outputs.logits ``` a" Example: ```python >>> from transformers import AutoProcessor, {model_class} >>> from datasets import load_dataset >>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation", trust_remote_code=True) >>> dataset = dataset.sort("id") >>> sampling_rate = dataset.features["audio"].sampling_rate >>> processor = AutoProcessor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> # audio file is decoded on the fly >>> inputs = processor(dataset[0]["audio"]["array"], sampling_rate=sampling_rate, return_tensors="tf") >>> outputs = model(**inputs) >>> last_hidden_states = outputs.last_hidden_state >>> list(last_hidden_states.shape) {expected_output} ``` aw Example: ```python >>> from transformers import AutoProcessor, {model_class} >>> from datasets import load_dataset >>> import tensorflow as tf >>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation", trust_remote_code=True) >>> dataset = dataset.sort("id") >>> sampling_rate = dataset.features["audio"].sampling_rate >>> processor = AutoProcessor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> # audio file is decoded on the fly >>> inputs = processor(dataset[0]["audio"]["array"], sampling_rate=sampling_rate, return_tensors="tf") >>> logits = model(**inputs).logits >>> predicted_ids = tf.math.argmax(logits, axis=-1) >>> # transcribe speech >>> transcription = processor.batch_decode(predicted_ids) >>> transcription[0] {expected_output} ``` ```python >>> inputs["labels"] = processor(text=dataset[0]["text"], return_tensors="tf").input_ids >>> # compute loss >>> loss = model(**inputs).loss >>> round(float(loss), 2) {expected_loss} ``` aq Example: ```python >>> from transformers import AutoImageProcessor, {model_class} >>> from datasets import load_dataset >>> dataset = load_dataset("huggingface/cats-image", trust_remote_code=True) >>> image = dataset["test"]["image"][0] >>> image_processor = AutoImageProcessor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = image_processor(image, return_tensors="tf") >>> outputs = model(**inputs) >>> last_hidden_states = outputs.last_hidden_state >>> list(last_hidden_states.shape) {expected_output} ``` a Example: ```python >>> from transformers import AutoImageProcessor, {model_class} >>> import tensorflow as tf >>> from datasets import load_dataset >>> dataset = load_dataset("huggingface/cats-image", trust_remote_code=True) >>> image = dataset["test"]["image"][0] >>> image_processor = AutoImageProcessor.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = image_processor(image, return_tensors="tf") >>> logits = model(**inputs).logits >>> # model predicts one of the 1000 ImageNet classes >>> predicted_label = int(tf.math.argmax(logits, axis=-1)) >>> print(model.config.id2label[predicted_label]) {expected_output} ``` ) r]r^r_r`rarbrcrdrerirjar Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="jax") >>> outputs = model(**inputs) >>> logits = outputs.logits ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> question, text = "Who was Jim Henson?", "Jim Henson was a nice puppet" >>> inputs = tokenizer(question, text, return_tensors="jax") >>> outputs = model(**inputs) >>> start_scores = outputs.start_logits >>> end_scores = outputs.end_logits ``` a} Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("The capital of France is {mask}.", return_tensors="jax") >>> outputs = model(**inputs) >>> logits = outputs.logits ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="jax") >>> outputs = model(**inputs) >>> last_hidden_states = outputs.last_hidden_state ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> prompt = "In Italy, pizza served in formal settings, such as at a restaurant, is presented unsliced." >>> choice0 = "It is eaten with a fork and a knife." >>> choice1 = "It is eaten while held in the hand." >>> encoding = tokenizer([prompt, prompt], [choice0, choice1], return_tensors="jax", padding=True) >>> outputs = model(**{{k: v[None, :] for k, v in encoding.items()}}) >>> logits = outputs.logits ``` a Example: ```python >>> from transformers import AutoTokenizer, {model_class} >>> tokenizer = AutoTokenizer.from_pretrained("{checkpoint}") >>> model = {model_class}.from_pretrained("{checkpoint}") >>> inputs = tokenizer("Hello, my dog is cute", return_tensors="np") >>> outputs = model(**inputs) >>> # retrieve logts for next token >>> next_token_logits = outputs.logits[:, -1] ``` )r]r^r_r`rarcrbcKsB|D]\}}|dur qd|d}td|dd|}q|S)zo Removes the lines testing an output with the doctest syntax in a code sample when it's set to `None`. N{}z\n([^\n]+)\n\s+z\nr>)itemsr9rC)r6kwargskeyvaluedoc_keyr r rfilter_outputs_from_example5s  rrz[MASK]) processor_class checkpointrVrLmaskqa_target_start_indexqa_target_end_index model_clsmodalityexpected_output expected_lossreal_checkpointrevisionc s* fdd}|S)Nc sfdur |jddn}|dddkrt}n |dddkr$t}nt}|   dd }d |vs.docstring_decoratorr )rVrLrr rrreplace_return_docstringssrcCs6tj|j|j|j|j|jd}t||}|j |_ |S)zReturns a copy of a function f.)nameargdefsclosure) types FunctionType__code__ __globals__rP __defaults__ __closure__ functoolsupdate_wrapper__kwdefaults__)fgr r r copy_funcs r)N)NN)7rrrr9r%rrrr7r8rTrSr=rHr\rPT_TOKEN_CLASSIFICATION_SAMPLEPT_QUESTION_ANSWERING_SAMPLE!PT_SEQUENCE_CLASSIFICATION_SAMPLEPT_MASKED_LM_SAMPLEPT_BASE_MODEL_SAMPLEPT_MULTIPLE_CHOICE_SAMPLEPT_CAUSAL_LM_SAMPLEPT_SPEECH_BASE_MODEL_SAMPLEPT_SPEECH_CTC_SAMPLEPT_SPEECH_SEQ_CLASS_SAMPLEPT_SPEECH_FRAME_CLASS_SAMPLEPT_SPEECH_XVECTOR_SAMPLEPT_VISION_BASE_MODEL_SAMPLEPT_VISION_SEQ_CLASS_SAMPLErTF_TOKEN_CLASSIFICATION_SAMPLETF_QUESTION_ANSWERING_SAMPLE!TF_SEQUENCE_CLASSIFICATION_SAMPLETF_MASKED_LM_SAMPLETF_BASE_MODEL_SAMPLETF_MULTIPLE_CHOICE_SAMPLETF_CAUSAL_LM_SAMPLETF_SPEECH_BASE_MODEL_SAMPLETF_SPEECH_CTC_SAMPLETF_VISION_BASE_MODEL_SAMPLETF_VISION_SEQ_CLASS_SAMPLEr FLAX_TOKEN_CLASSIFICATION_SAMPLEFLAX_QUESTION_ANSWERING_SAMPLE#FLAX_SEQUENCE_CLASSIFICATION_SAMPLEFLAX_MASKED_LM_SAMPLEFLAX_BASE_MODEL_SAMPLEFLAX_MULTIPLE_CHOICE_SAMPLEFLAX_CAUSAL_LM_SAMPLErrrrrrr r r rs  $   + !": #$#!#!$  c