repo_id
stringlengths 15
89
| file_path
stringlengths 27
180
| content
stringlengths 1
2.23M
| __index_level_0__
int64 0
0
|
|---|---|---|---|
hf_public_repos/transformers/tests
|
hf_public_repos/transformers/tests/fixtures/input.txt
|
Who was Jim Henson ? ||| Jim Henson was a puppeteer
| 0
|
hf_public_repos/transformers/tests
|
hf_public_repos/transformers/tests/fixtures/test_entity_vocab.json
|
{"[MASK]": 0, "[UNK]": 1, "[PAD]": 2, "DUMMY": 3, "DUMMY2": 4, "[MASK2]": 5}
| 0
|
hf_public_repos/transformers/tests
|
hf_public_repos/transformers/tests/fixtures/sample_text.txt
|
This text is included to make sure Unicode is handled properly: ๅๅ ๅๅๅบแดตแดบแตแตเฆเฆเฆเฆกเฆฃเฆค
Text should be one-sentence-per-line, with empty lines between documents.
This sample text is public domain and was randomly selected from Project Guttenberg.
The rain had only ceased with the gray streaks of morning at Blazing Star, and the settlement awoke to a moral sense of cleanliness, and the finding of forgotten knives, tin cups, and smaller camp utensils, where the heavy showers had washed away the debris and dust heaps before the cabin doors.
Indeed, it was recorded in Blazing Star that a fortunate early riser had once picked up on the highway a solid chunk of gold quartz which the rain had freed from its incumbering soil, and washed into immediate and glittering popularity.
Possibly this may have been the reason why early risers in that locality, during the rainy season, adopted a thoughtful habit of body, and seldom lifted their eyes to the rifted or india-ink washed skies above them.
"Cass" Beard had risen early that morning, but not with a view to discovery.
A leak in his cabin roof,--quite consistent with his careless, improvident habits,--had roused him at 4 A. M., with a flooded "bunk" and wet blankets.
The chips from his wood pile refused to kindle a fire to dry his bed-clothes, and he had recourse to a more provident neighbor's to supply the deficiency.
This was nearly opposite.
Mr. Cassius crossed the highway, and stopped suddenly.
Something glittered in the nearest red pool before him.
Gold, surely!
But, wonderful to relate, not an irregular, shapeless fragment of crude ore, fresh from Nature's crucible, but a bit of jeweler's handicraft in the form of a plain gold ring.
Looking at it more attentively, he saw that it bore the inscription, "May to Cass."
Like most of his fellow gold-seekers, Cass was superstitious.
The fountain of classic wisdom, Hypatia herself.
As the ancient sage--the name is unimportant to a monk--pumped water nightly that he might study by day, so I, the guardian of cloaks and parasols, at the sacred doors of her lecture-room, imbibe celestial knowledge.
From my youth I felt in me a soul above the matter-entangled herd.
She revealed to me the glorious fact, that I am a spark of Divinity itself.
A fallen star, I am, sir!' continued he, pensively, stroking his lean stomach--'a fallen star!--fallen, if the dignity of philosophy will allow of the simile, among the hogs of the lower world--indeed, even into the hog-bucket itself. Well, after all, I will show you the way to the Archbishop's.
There is a philosophic pleasure in opening one's treasures to the modest young.
Perhaps you will assist me by carrying this basket of fruit?' And the little man jumped up, put his basket on Philammon's head, and trotted off up a neighbouring street.
Philammon followed, half contemptuous, half wondering at what this philosophy might be, which could feed the self-conceit of anything so abject as his ragged little apish guide;
but the novel roar and whirl of the street, the perpetual stream of busy faces, the line of curricles, palanquins, laden asses, camels, elephants, which met and passed him, and squeezed him up steps and into doorways, as they threaded their way through the great Moon-gate into the ample street beyond, drove everything from his mind but wondering curiosity, and a vague, helpless dread of that great living wilderness, more terrible than any dead wilderness of sand which he had left behind.
Already he longed for the repose, the silence of the Laura--for faces which knew him and smiled upon him; but it was too late to turn back now.
His guide held on for more than a mile up the great main street, crossed in the centre of the city, at right angles, by one equally magnificent, at each end of which, miles away, appeared, dim and distant over the heads of the living stream of passengers, the yellow sand-hills of the desert;
while at the end of the vista in front of them gleamed the blue harbour, through a network of countless masts.
At last they reached the quay at the opposite end of the street;
and there burst on Philammon's astonished eyes a vast semicircle of blue sea, ringed with palaces and towers.
He stopped involuntarily; and his little guide stopped also, and looked askance at the young monk, to watch the effect which that grand panorama should produce on him.
| 0
|
hf_public_repos/transformers/tests
|
hf_public_repos/transformers/tests/fixtures/dummy-config.json
|
{
"model_type": "roberta"
}
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/MRPC/train.csv
|
label,sentence1,sentence2
equivalent,He said the foodservice pie business doesn 't fit the company 's long-term growth strategy .,""" The foodservice pie business does not fit our long-term growth strategy ."
not_equivalent,Magnarelli said Racicot hated the Iraqi regime and looked forward to using his long years of training in the war .,"His wife said he was "" 100 percent behind George Bush "" and looked forward to using his years of training in the war ."
not_equivalent,"The dollar was at 116.92 yen against the yen , flat on the session , and at 1.2891 against the Swiss franc , also flat .","The dollar was at 116.78 yen JPY = , virtually flat on the session , and at 1.2871 against the Swiss franc CHF = , down 0.1 percent ."
equivalent,The AFL-CIO is waiting until October to decide if it will endorse a candidate .,The AFL-CIO announced Wednesday that it will decide in October whether to endorse a candidate before the primaries .
not_equivalent,No dates have been set for the civil or the criminal trial .,"No dates have been set for the criminal or civil cases , but Shanley has pleaded not guilty ."
equivalent,Wal-Mart said it would check all of its million-plus domestic workers to ensure they were legally employed .,It has also said it would review all of its domestic employees more than 1 million to ensure they have legal status .
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/MRPC/dev.tsv
|
๏ปฟQuality #1 ID #2 ID #1 String #2 String
1 1355540 1355592 He said the foodservice pie business doesn 't fit the company 's long-term growth strategy . " The foodservice pie business does not fit our long-term growth strategy .
0 2029631 2029565 Magnarelli said Racicot hated the Iraqi regime and looked forward to using his long years of training in the war . His wife said he was " 100 percent behind George Bush " and looked forward to using his years of training in the war .
0 487993 487952 The dollar was at 116.92 yen against the yen , flat on the session , and at 1.2891 against the Swiss franc , also flat . The dollar was at 116.78 yen JPY = , virtually flat on the session , and at 1.2871 against the Swiss franc CHF = , down 0.1 percent .
1 1989515 1989458 The AFL-CIO is waiting until October to decide if it will endorse a candidate . The AFL-CIO announced Wednesday that it will decide in October whether to endorse a candidate before the primaries .
0 1783137 1782659 No dates have been set for the civil or the criminal trial . No dates have been set for the criminal or civil cases , but Shanley has pleaded not guilty .
1 3039165 3039036 Wal-Mart said it would check all of its million-plus domestic workers to ensure they were legally employed . It has also said it would review all of its domestic employees more than 1 million to ensure they have legal status .
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/MRPC/train.tsv
|
๏ปฟQuality #1 ID #2 ID #1 String #2 String
1 1355540 1355592 He said the foodservice pie business doesn 't fit the company 's long-term growth strategy . " The foodservice pie business does not fit our long-term growth strategy .
0 2029631 2029565 Magnarelli said Racicot hated the Iraqi regime and looked forward to using his long years of training in the war . His wife said he was " 100 percent behind George Bush " and looked forward to using his years of training in the war .
0 487993 487952 The dollar was at 116.92 yen against the yen , flat on the session , and at 1.2891 against the Swiss franc , also flat . The dollar was at 116.78 yen JPY = , virtually flat on the session , and at 1.2871 against the Swiss franc CHF = , down 0.1 percent .
1 1989515 1989458 The AFL-CIO is waiting until October to decide if it will endorse a candidate . The AFL-CIO announced Wednesday that it will decide in October whether to endorse a candidate before the primaries .
0 1783137 1782659 No dates have been set for the civil or the criminal trial . No dates have been set for the criminal or civil cases , but Shanley has pleaded not guilty .
1 3039165 3039036 Wal-Mart said it would check all of its million-plus domestic workers to ensure they were legally employed . It has also said it would review all of its domestic employees more than 1 million to ensure they have legal status .
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/MRPC/dev.csv
|
label,sentence1,sentence2
equivalent,He said the foodservice pie business doesn 't fit the company 's long-term growth strategy .,""" The foodservice pie business does not fit our long-term growth strategy ."
not_equivalent,Magnarelli said Racicot hated the Iraqi regime and looked forward to using his long years of training in the war .,"His wife said he was "" 100 percent behind George Bush "" and looked forward to using his years of training in the war ."
not_equivalent,"The dollar was at 116.92 yen against the yen , flat on the session , and at 1.2891 against the Swiss franc , also flat .","The dollar was at 116.78 yen JPY = , virtually flat on the session , and at 1.2871 against the Swiss franc CHF = , down 0.1 percent ."
equivalent,The AFL-CIO is waiting until October to decide if it will endorse a candidate .,The AFL-CIO announced Wednesday that it will decide in October whether to endorse a candidate before the primaries .
not_equivalent,No dates have been set for the civil or the criminal trial .,"No dates have been set for the criminal or civil cases , but Shanley has pleaded not guilty ."
equivalent,Wal-Mart said it would check all of its million-plus domestic workers to ensure they were legally employed .,It has also said it would review all of its domestic employees more than 1 million to ensure they have legal status .
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/wiki_text/wiki_00
|
<doc id="12" url="https://en.wikipedia.org/wiki?curid=12" title="Anarchism">
Anarchism
Anarchism is a political philosophy and movement that rejects all involuntary, coercive forms of hierarchy. It radically calls for the abolition of the state which it holds to be undesirable, unnecessary, and harmful.
The history of anarchism stretches back to prehistory, when humans lived in anarchistic societies long before the establishment of formal states, realms or empires. With the rise of organised hierarchical bodies, skepticism toward authority also rose, but it was not until the 19th century that a self-conscious political movement emerged. During the latter half of the 19th and the first decades of the 20th century, the anarchist movement flourished in most parts of the world and had a significant role in worker's struggles for emancipation. Various anarchist schools of thought formed during this period.
Anarchists took part in several revolutions, most notably in the Spanish Civil War, where they were crushed along with the alliance to restore the Second Republic by the fascist forces of the Nationalist faction and its foreign allies in Nazi Germany, Fascist Italy, Portuguese Dictatorship and the Catholic Church in 1939, marking the end of the classical era of anarchism. In the last decades of the 20th century and into the 21st century, the anarchist movement has been resurgent once more.
Anarchism employs various tactics in order to meet its ideal ends; these can be broadly separated into revolutionary and evolutionary tactics. There is significant overlap between the two, which are merely descriptive. Revolutionary tactics aim to bring down authority and state, and have taken a violent turn in the past. Evolutionary tactics aim to prefigure what an anarchist society would be like. Anarchist thought, criticism, and praxis has played a part in diverse areas of human society.
The etymological origin of "anarchism" is from the Ancient Greek "anarkhia", meaning "without a ruler", composed of the prefix "an-" (i.e. "without") and the word "arkhos" (i.e. "leader" or "ruler"). The suffix "-ism" denotes the ideological current that favours anarchy. "Anarchism" appears in English from 1642 as "anarchisme" and "anarchy" from 1539. Various factions within the French Revolution labelled their opponents as "anarchists", although few such accused shared many views with later anarchists. Many revolutionaries of the 19th century such as William Godwin (1756โ1836) and Wilhelm Weitling (1808โ1871) would contribute to the anarchist doctrines of the next generation, but they did not use "anarchist" or "anarchism" in describing themselves or their beliefs.
The first political philosopher to call himself an "anarchist" () was Pierre-Joseph Proudhon (1809โ1865), marking the formal birth of anarchism in the mid-19th century. Since the 1890s and beginning in France, "libertarianism" has often been used as a synonym for anarchism and its use as a synonym is still common outside the United States. On the other hand, some use "libertarianism" to refer to individualistic free-market philosophy only, referring to free-market anarchism as "libertarian anarchism".
While opposition to the state is central to anarchist thought, defining anarchism is not an easy task as there is a lot of discussion among scholars and anarchists on the matter and various currents perceive anarchism slightly differently. Hence, it might be true to say that anarchism is a cluster of political philosophies opposing authority and hierarchical organization (including the state, capitalism, nationalism and all associated institutions) in the conduct of all human relations in favour of a society based on voluntary association, on freedom and on decentralisation, but this definition has the same shortcomings as the definition based on etymology (which is simply a negation of a ruler), or based on anti-statism (anarchism is much more than that) or even the anti-authoritarian (which is an "a posteriori" conclusion). Nonetheless, major elements of the definition of anarchism include the following:
During the prehistoric era of mankind, an established authority did not exist. It was after the creation of towns and cities that institutions of authority were established and anarchistic ideas espoused as a reaction. Most notable precursors to anarchism in the ancient world were in China and Greece. In China, philosophical anarchism (i.e. the discussion on the legitimacy of the state) was delineated by Taoist philosophers Zhuang Zhou and Laozi.
Likewise, anarchic attitudes were articulated by tragedians and philosophers in Greece. Aeschylus and Sophocles used the myth of Antigone to illustrate the conflict between rules set by the state and personal autonomy. Socrates questioned Athenian authorities constantly and insisted to the right of individual freedom of consciousness. Cynics dismissed human law ("nomos") and associated authorities while trying to live according to nature ("physis"). Stoics were supportive of a society based on unofficial and friendly relations among its citizens without the presence of a state.
During the Middle Ages, there was no anarchistic activity except some ascetic religious movements in the Muslim world or in Christian Europe. This kind of tradition later gave birth to religious anarchism. In the Sasanian Empire, Mazdak called for an egalitarian society and the abolition of monarchy, only to be soon executed by Emperor Kavad I.
In Basra, religious sects preached against the state. In Europe, various sects developed anti-state and libertarian tendencies. Libertarian ideas further emerged during the Renaissance with the spread of reasoning and humanism through Europe. Novelists fictionalised ideal societies that were based not on coercion but voluntarism. The Enlightenment further pushed towards anarchism with the optimism for social progress.
During the French Revolution, partisan groups such as the Enragรฉs and the saw a turning point in the fermentation of anti-state and federalist sentiments. The first anarchist currents developed throughout the 18th centuryโWilliam Godwin espoused philosophical anarchism in England, morally delegitimizing the state, Max Stirner's thinking paved the way to individualism, and Pierre-Joseph Proudhon's theory of mutualism found fertile soil in France. This era of classical anarchism lasted until the end of the Spanish Civil War of 1936 and is considered the golden age of anarchism.
Drawing from mutualism, Mikhail Bakunin founded collectivist anarchism and entered the International Workingmen's Association, a class worker union later known as the First International that formed in 1864 to unite diverse revolutionary currents. The International became a significant political force, with Karl Marx being a leading figure and a member of its General Council. Bakunin's faction (the Jura Federation) and Proudhon's followers (the mutualists) opposed Marxist state socialism, advocating political abstentionism and small property holdings. After bitter disputes, the Bakuninists were expelled from the International by the Marxists at the 1872 Hague Congress. Bakunin famously predicted that if revolutionaries gained power by Marx's terms, they would end up the new tyrants of workers. After being expelled, anarchists formed the St. Imier International. Under the influence of Peter Kropotkin, a Russian philosopher and scientist, anarcho-communism overlapped with collectivism. Anarcho-communists, who drew inspiration from the 1871 Paris Commune, advocated for free federation and for the distribution of goods according to one's needs.
At the turn of the century, anarchism had spread all over the world. In China, small groups of students imported the humanistic pro-science version of anarcho-communism. Tokyo was a hotspot for rebellious youth from countries of the far east, travelling to the Japanese capital to study. In Latin America, Argentina was a stronghold for anarcho-syndicalism, where it became the most prominent left-wing ideology. During this time, a minority of anarchists adopted tactics of revolutionary political violence. This strategy became known as propaganda of the deed. The dismemberment of the French socialist movement into many groups, and the execution and exile of many Communards to penal colonies following the suppression of the Paris Commune, favoured individualist political expression and acts. Even though many anarchists distanced themselves from these terrorist acts, infamy came upon the movement. Illegalism was another strategy which some anarchists adopted during this period.
Anarchists enthusiastically participated in the Russian Revolutionโdespite concernsโin opposition to the Whites. However, they met harsh suppression after the Bolshevik government was stabilized. Several anarchists from Petrograd and Moscow fled to Ukraine, notably leading to the Kronstadt rebellion and Nestor Makhno's struggle in the Free Territory. With the anarchists being crushed in Russia, two new antithetical currents emerged, namely platformism and synthesis anarchism. The former sought to create a coherent group that would push for revolution while the latter were against anything that would resemble a political party. Seeing the victories of the Bolsheviks in the October Revolution and the resulting Russian Civil War, many workers and activists turned to communist parties, which grew at the expense of anarchism and other socialist movements. In France and the United States, members of major syndicalist movements, the General Confederation of Labour and Industrial Workers of the World, left their organisations and joined the Communist International.
In the Spanish Civil War, anarchists and syndicalists (CNT and FAI) once again allied themselves with various currents of leftists. A long tradition of Spanish anarchism led to anarchists playing a pivotal role in the war. In response to the army rebellion, an anarchist-inspired movement of peasants and workers, supported by armed militias, took control of Barcelona and of large areas of rural Spain, where they collectivised the land. The Soviet Union provided some limited assistance at the beginning of the war, but the result was a bitter fight among communists and anarchists at a series of events named May Days as Joseph Stalin tried to seize control of the Republicans.
At the end of World War II, the anarchist movement was severely weakened. However, the 1960s witnessed a revival of anarchism likely caused by a perceived failure of MarxismโLeninism and tensions built by the Cold War. During this time, anarchism took root in other movements critical towards both the state and capitalism, such as the anti-nuclear, environmental and pacifist movements, the New Left, and the counterculture of the 1960s. Anarchism became associated with punk subculture, as exemplified by bands such as Crass and the Sex Pistols, and the established feminist tendencies of anarcha-feminism returned with vigour during the second wave of feminism.
Around the turn of the 21st century, anarchism grew in popularity and influence within anti-war, anti-capitalist, and anti-globalisation movements. Anarchists became known for their involvement in protests against the World Trade Organization, the Group of Eight and the World Economic Forum. During the protests, "ad hoc" leaderless anonymous cadres known as black blocs engaged in rioting, property destruction, and violent confrontations with the police. Other organisational tactics pioneered in this time include security culture, affinity groups, and the use of decentralised technologies such as the internet. A significant event of this period was the confrontations at the WTO conference in Seattle in 1999. Anarchist ideas have been influential in the development of the Zapatistas in Mexico and the Democratic Federation of Northern Syria, more commonly known as Rojava, a "de facto" autonomous region in northern Syria.
Anarchist schools of thought have been generally grouped into two main historical traditions, social anarchism and individualist anarchism, owing to their different origins, values and evolution. The individualist current emphasises negative liberty in opposing restraints upon the free individual, while the social current emphasises positive liberty in aiming to achieve the free potential of society through equality and social ownership. In a chronological sense, anarchism can be segmented by the classical currents of the late 19th century, and the post-classical currents (such as anarcha-feminism, green anarchism and post-anarchism) developed thereafter.
Beyond the specific factions of anarchist movements which constitute political anarchism lies philosophical anarchism, which holds that the state lacks moral legitimacy, without necessarily accepting the imperative of revolution to eliminate it. A component especially of individualist anarchism, philosophical anarchism may tolerate the existence of a minimal state, but argues that citizens have no moral obligation to obey government when it conflicts with individual autonomy. Anarchism pays significant attention to moral arguments since ethics have a central role in anarchist philosophy.
One reaction against sectarianism within the anarchist milieu was anarchism without adjectives, a call for toleration and unity among anarchists first adopted by Fernando Tarrida del Mรกrmol in 1889 in response to the bitter debates of anarchist theory at the time. Despite separation, the various anarchist schools of thought are not seen as distinct entities, but as tendencies that intermingle.
Anarchism is usually placed on the far-left of the political spectrum. Much of its economics and legal philosophy reflect anti-authoritarian, anti-statist, and libertarian interpretations of the radical left-wing and socialist politics of collectivism, communism, individualism, mutualism, and syndicalism, among other libertarian socialist economic theories. As anarchism does not offer a fixed body of doctrine from a single particular worldview, many anarchist types and traditions exist, and varieties of anarchy diverge widely.
Inceptive currents among classical anarchist currents were mutualism and individualism. They were followed by the major currents of social anarchism (collectivist, communist, and syndicalist). They differ on organizational and economic aspects of their ideal society.
Mutualism is an 18th-century economic theory that was developed into anarchist theory by Pierre-Joseph Proudhon. Its aims include reciprocity, free association, voluntary contract, federation, and credit and currency reform that would be regulated by a bank of the people. Mutualism has been retrospectively characterised as ideologically situated between individualist and collectivist forms of anarchism. Proudhon first characterised his goal as a "third form of society, the synthesis of communism and property".
Collectivist anarchism, also known as anarchist collectivism or anarcho-collectivism, is a revolutionary socialist form of anarchism commonly associated with Mikhail Bakunin. Collectivist anarchists advocate collective ownership of the means of production, theorised to be achieved through violent revolution, and that workers be paid according to time worked, rather than goods being distributed according to need as in communism. Collectivist anarchism arose alongside Marxism, but rejected the dictatorship of the proletariat despite the stated Marxist goal of a collectivist stateless society. Anarcho-communism, also known as anarchist-communism, communist anarchism, and libertarian communism, is a theory of anarchism that advocates a communist society with common ownership of the means of production, direct democracy, and a horizontal network of voluntary associations and workers' councils with production and consumption based on the guiding principle: "From each according to his ability, to each according to his need". Anarcho-communism developed from radical socialist currents after the French Revolution, but it was first formulated as such in the Italian section of the First International. It was later expanded upon in the theoretical work of Peter Kropotkin.
Anarcho-syndicalism, also referred to as revolutionary syndicalism, is a branch of anarchism that views labour syndicates as a potential force for revolutionary social change, replacing capitalism and the state with a new society democratically self-managed by workers. The basic principles of anarcho-syndicalism are workers' solidarity, direct action, and workers' self-management.
Individualist anarchism refers to several traditions of thought within the anarchist movement that emphasise the individual and their will over any kinds of external determinants. Early influences on individualist forms of anarchism include William Godwin, Max Stirner and Henry David Thoreau. Through many countries, individualist anarchism attracted a small yet diverse following of Bohemian artists and intellectuals as well as young anarchist outlaws in what became known as illegalism and individual reclamation.
Anarchist principles undergird contemporary radical social movements of the left. Interest in the anarchist movement developed alongside momentum in the anti-globalization movement, whose leading activist networks were anarchist in orientation. As the movement shaped 21st century radicalism, wider embrace of anarchist principles signaled a revival of interest. Contemporary news coverage which emphasizes black bloc demonstrations has reinforced anarchism's historical association with chaos and violence, although its publicity has also led more scholars to engage with the anarchist movement. Anarchism has continued to generate many philosophies and movementsโat times eclectic, drawing upon various sources, and syncretic, combining disparate concepts to create new philosophical approaches. The anti-capitalist tradition of classical anarchism has remained prominent within contemporary currents.
Various anarchist groups, tendencies, and schools of thought exist today, making it difficult to describe contemporary anarchist movement. While theorists and activists have established "relatively stable constellations of anarchist principles", there is no consensus on which principles are core. As a result, commentators describe multiple "anarchisms" (rather than a singular "anarchism") in which common principles are shared between schools of anarchism while each group prioritizes those principles differently. For example, gender equality can be a common principle but ranks as a higher priority to anarcha-feminists than anarchist communists. Anarchists are generally committed against coercive authority in all forms, namely "all centralized and hierarchical forms of government (e.g., monarchy, representative democracy, state socialism, etc.), economic class systems (e.g., capitalism, Bolshevism, feudalism, slavery, etc.), autocratic religions (e.g., fundamentalist Islam, Roman Catholicism, etc.), patriarchy, heterosexism, white supremacy, and imperialism". However, anarchist schools disagree on the methods by which these forms should be opposed.
Anarchists' tactics take various forms but in general serve two major goalsโfirst, to oppose the Establishment; and second, to promote anarchist ethics and reflect an anarchist vision of society, illustrating the unity of means and ends. A broad categorization can be made between aims to destroy oppressive states and institutions by revolutionary means, and aims to change society through evolutionary means. Evolutionary tactics reject violence and take a gradual approach to anarchist aims, though there is significant overlap between the two.
Anarchist tactics have shifted during the course of the last century. Anarchists during the early 20th century focused more on strikes and militancy, while contemporary anarchists use a broader array of approaches.
During the classical era, anarchists had a militant tendency. Not only did they confront state armed forces (as in Spain and Ukraine) but some of them also employed terrorism as propaganda of the deed. Assassination attempts were carried out against heads of state, some of which were successful. Anarchists also took part in revolutions. Anarchist perspectives towards violence have always been perplexing and controversial. On one hand, anarcho-pacifists point out the unity of means and ends. On the other hand, other anarchist groups advocate direct action, a tactic which can include acts of sabotage or even acts of terrorism. This attitude was quite prominent a century ago; seeing the state as a tyrant, some anarchists believed that they had every right to oppose its oppression by any means possible. Emma Goldman and Errico Malatesta, who were proponents of limited use of violence, argued that violence is merely a reaction to state violence as a necessary evil.
Anarchists took an active role in strikes, although they tended to be antipathetic to formal syndicalism, seeing it as reformist. They saw it as a part of the movement which sought to overthrow the state and capitalism. Anarchists also reinforced their propaganda within the arts, some of whom practiced nudism. They also built communities which were based on friendship. They were also involved in the press.
In the current era, Italian anarchist Alfredo Bonanno, a proponent of insurrectionary anarchism, has reinstated the debate on violence by rejecting the nonviolence tactic adopted since the late 19th century by Kropotkin and other prominent anarchists afterwards. Both Bonanno and the French group The Invisible Committee advocate for small, informal affiliation groups, where each member is responsible for their own actions but works together to bring down oppression utilizing sabotage and other violent means against state, capitalism and other enemies. Members of The Invisible Committee were arrested in 2008 on various charges, terrorism included.
Overall, today's anarchists are much less violent and militant than their ideological ancestors. They mostly engage in confronting the police during demonstrations and riots, especially in countries like Canada, Mexico or Greece. ฮilitant black bloc protest groups are known for clashing with the police. However, anarchists not only clash with state operators; they also engage in the struggle against fascists and racists, taking anti-fascist action and mobilizing to prevent hate rallies from happening.
Anarchists commonly employ direct action. This can take the form of disrupting and protesting against unjust hierarchy, or the form of self-managing their lives through the creation of counter-institutions such as communes and non-hierarchical collectives. Often, decision-making is handled in an anti-authoritarian way, with everyone having equal say in each decision, an approach known as horizontalism. Contemporary-era anarchists have been engaging with various grassroots movements that are not explicitly anarchist but are more or less based on horizontalism, respecting personal autonomy, and participating in mass activism such as strikes and demonstrations. The newly coined term "small-a anarchism", in contrast with the "big-A anarchism" of the classical era, signals their tendency not to base their thoughts and actions on classical-era anarchism or to refer to Kropotkin or Proudhon to justify their opinions. They would rather base their thought and praxis on their own experience, which they will later theorize.
The decision-making process of small affinity anarchist groups play a significant tactical role. Anarchists have employed various methods in order to build a rough consensus among members of their group, without the need of a leader or a leading group. One way is for an individual from the group to play the role of facilitator to help achieve a consensus without taking part in the discussion themselves or promoting a specific point. Minorities usually accept rough consensus, except when they feel the proposal contradicts anarchist goals, values, or ethics. Anarchists usually form small groups (5โ20 individuals) to enhance autonomy and friendships among their members. These kind of groups more often than not interconnect with each other, forming larger networks. Anarchists still support and participate in strikes, especially wildcat strikes; these are leaderless strikes not organised centrally by a syndicate.
Anarchists have gone online to spread their message. As in the past, newspapers and journals are used; however, because of distributional and other difficulties, anarchists have found it easier to create websites, hosting electronic libraries and other portals. Anarchists were also involved in developing various software that are available for free. The way these hacktivists work to develop and distribute resembles the anarchist ideals, especially when it comes to preserving user's privacy from state surveillance.
Anarchists organize themselves to squat and reclaim public spaces. During important events such as protests and when spaces are being occupied, they are often called Temporary Autonomous Zones (TAZ), spaces where surrealism, poetry and art are blended to display the anarchist ideal. As seen by anarchists, squatting is a way to regain urban space from the capitalist market, serving pragmatical needs, and is also seen an exemplary direct action. Acquiring space enables anarchists to experiment with their ideas and build social bonds. Adding up these tactics, and having in mind that not all anarchists share the same attitudes towards them, along with various forms of protesting at highly symbolic events, make up a carnivalesque atmosphere that is part of contemporary anarchist vividity.
As anarchism is a philosophy that embodies many diverse attitudes, tendencies, and schools of thought, and disagreement over questions of values, ideology, and tactics is common, its diversity has led to widely different uses of identical terms among different anarchist traditions, which has created a number of definitional concerns in anarchist theory. For instance, the compatibility of capitalism, nationalism and religion with anarchism is widely disputed. Similarly, anarchism enjoys complex relationships with ideologies such as Marxism, communism, collectivism and trade unionism. Anarchists may be motivated by humanism, divine authority, enlightened self-interest, veganism, or any number of alternative ethical doctrines. Phenomena such as civilisation, technology (e.g. within anarcho-primitivism) and the democratic process may be sharply criticised within some anarchist tendencies and simultaneously lauded in others.
Gender and sexuality carry along them dynamics of hierarchy; anarchism is obliged to address, analyse and oppose the suppression of one's autonomy because of the dynamics that gender roles traditionally impose.
A historical current that arose and flourished during 1890 and 1920 within anarchism was free love; in contemporary anarchism, this current survives as a tendency to support polyamory and queer anarchism. Free love advocates were against marriage, which they saw as a way of men imposing authority over women, largely because marriage law greatly favoured the power of men. The notion of free love, though, was much broader; it included critique of the established order that limited women's sexual freedom and pleasure. Such free love movements contributed to the establishment of communal houses, where large groups of travelers, anarchists, and other activists slept in beds together. Free love had roots both in Europe and the United States. Some anarchists, however, struggled with the jealousy that arose from free love. Anarchist feminists were advocates of free love, against marriage, were pro-choice (utilizing a contemporary term) and had a likewise agenda. Anarchist and non-anarchist feminists differed on suffrage, but were nonetheless supportive of one another.
During the second half of the 20th century, anarchism intermingled with the second wave of feminism, radicalizing some currents of the feminist movement (and being influenced as well). By the latest decades of the 20th century, anarchists and feminists were advocating for the rights and autonomy of women, gays, queers and other marginalized groups, with some feminist thinkers suggesting a fusion of the two currents. With the third wave of feminism, sexual identity and compulsory heterosexuality became a subject of study for anarchists, which yielded a post-structuralist critique of sexual normality. However, some anarchists distanced themselves from this line of thinking, suggesting that it leaned towards individualism and was, therefore, dropping the cause of social liberation.
The interest of anarchists in education stretches back to the first emergence of classical anarchism. Anarchists consider 'proper' education, which sets the foundations of the future autonomy of the individual and the society, to be an act of mutual aid. Anarchist writers such as Willian Godwin and Max Stirner attacked both state education and private education as another means by which the ruling class replicate their privileges.
In 1901, Catalan anarchist and free thinker Francisco Ferrer established the Escuela Moderna in Barcelona as an opposition to the established education system, which was dictated largely by the Catholic Church. Ferrer's approach was secular, rejecting both state and church involvement in the educational process, and gave pupils large amounts of autonomy in planning their work and attendance. Ferrer aimed to educate the working class and explicitly sought to foster class consciousness among students. The school closed after constant harassment by the state and Ferrer was later arrested. His ideas, however, formed the inspiration for a series of modern schools around the world. Christian anarchist Leo Tolstoy also established a similar school, with its founding principle, according to Tolstoy, being that "for education to be effective it had to be free". In a similar token, A. S. Neill founding what became Summerhill School in 1921, also declaring being free from coercion.
Anarchist education is based largely on the idea that a child's right to develop freely, without manipulation, ought to be respected, and that rationality will lead children to morally good conclusions. However, there has been little consensus among anarchist figures as to what constitutes manipulation; Ferrer, for example, believed that moral indoctrination was necessary and explicitly taught pupils that equality, liberty, and social justice were not possible under capitalism (along with other critiques of nationalism and government).
Late 20th century and contemporary anarchist writers (such as Colin Ward, Herbert Read and Paul Goodman) intensified and expanded the anarchist critique of state education, largely focusing on the need for a system that focuses on children's creativity rather than on their ability to attain a career or participate in consumer society. Contemporary anarchists, such as Colin Ward, have further argued that state education serves to perpetuate socio-economic inequality.
While few anarchist education institutions have survived to the modern day, major tenets of anarchist schools, such as respect for child autonomy and relying on reasoning rather than indoctrination as a teaching method, have spread among mainstream educational institutions.
Objection to the state and its institutions is a "sine qua non" of anarchism. Anarchists consider the state as a tool of domination and believe it to be illegitimate regardless of its political tendencies. Instead of people being able to control the aspects of their life, major decisions are taken by a small elite. Authority ultimately rests solely on power, regardless of whether that power is open or transparent, as it still has the ability to coerce people. Another anarchist argument against states is that the people constituting a government, even the most altruistic among officials, will unavoidably seek to gain more power, leading to corruption. Anarchists consider the idea that the state is the collective will of the people to be an unachievable fiction, due to the fact that the ruling class is distinct from the rest of society.
The connection between anarchism and art was quite profound during the classical era of anarchism, especially among artistic currents that were developing during that era, such as futurists, surrealists, and others, while in literature anarchism was mostly associated with the New Apocalyptics and the Neo-romanticism movement. In music, anarchism has been associated with music scenes such as Punk. Anarchists such as Leo Tolstoy and Herbert Read argued that the border between the artist and the non-artist, what separates art from a daily act, is a construct produced by the alienation caused by capitalism, and it prevents humans from living a joyful life.
Other anarchists advocated for or used art as a means to achieve anarchist ends. In his book Breaking the Spell: A History of Anarchist Filmmakers, Videotape Guerrillas, and Digital Ninjas Chris Robรฉ claims that "anarchist-inflected practices have increasingly structured movement-based video activism."
Three overlapping properties made art useful to anarchists: It could depict a critique of existing society and hierarchies; it could serve as a prefigurative tool to reflect the anarchist ideal society, and also it could turn into a means of direct action, in protests for example. As it appeals to both emotion and reason, art could appeal to the "whole human" and have a powerful effect.
Philosophy lecturer Andrew G. Fiala has listed five main arguments against anarchism. Firstly, he notes that anarchism is related to violence and destruction, not only in the pragmatic world (i.e. at protests) but in the world of ethics as well. The second argument is that it is impossible for a society to function without a state or something like a state, acting to protect citizens from criminality. Fiala takes "Leviathan" from Thomas Hobbes and the night-watchman state from philosopher Robert Nozick as examples. Thirdly, anarchism is evaluated as unfeasible or utopian since the state can not be defeated practically; this line of arguments most often calls for political action within the system to reform it. The fourth argument is that anarchism is self-contradictory since while it advocates for no-one to "archiei", if accepted by the many, then anarchism will turn into the ruling political theory. In this line of criticism also comes the self contradiction that anarchist calls for collective action while anarchism endorses the autonomy of the individual and hence no collective action can be taken. Lastly, Fiala mentions a critique towards philosophical anarchism, of being ineffective (all talk and thoughts) and in the meantime capitalism and bourgeois class remains strong.
Philosophical anarchism has met the criticism of members of academia, following the release of pro-anarchist books such as A. John Simmons' "Moral Principles and Political Obligations" (1979). Law professor William A. Edmundson authored an essay arguing against three major philosophical anarchist principles, which he finds fallacious; Edmundson claims that while the individual does not owe a normal state a duty of obedience, this does not imply that anarchism is the inevitable conclusion, and the state is still morally legitimate.
</doc>
<doc id="25" url="https://en.wikipedia.org/wiki?curid=25" title="Autism">
Autism
Autism is a developmental disorder characterized by difficulties with social interaction and communication, and by restricted and repetitive behavior. Parents often notice signs during the first three years of their child's life. These signs often develop gradually, though some children with autism experience worsening in their communication and social skills after reaching developmental milestones at a normal pace.
Autism is associated with a combination of genetic and environmental factors. Risk factors during pregnancy include certain infections, such as rubella, toxins including valproic acid, alcohol, cocaine, pesticides, lead, and air pollution, fetal growth restriction, and autoimmune diseases. Controversies surround other proposed environmental causes; for example, the vaccine hypothesis, which has been disproven. Autism affects information processing in the brain and how nerve cells and their synapses connect and organize; how this occurs is not well understood. The Diagnostic and Statistical Manual of Mental Disorders (DSM-5), combines autism and less severe forms of the condition, including Asperger syndrome and pervasive developmental disorder not otherwise specified (PDD-NOS) into the diagnosis of autism spectrum disorder (ASD).
Early behavioral interventions or speech therapy can help children with autism gain self-care, social, and communication skills. Although there is no known cure, there have been cases of children who recovered. Some autistic adults are unable to live independently. An autistic culture has developed, with some individuals seeking a cure and others believing autism should be accepted as a difference to be accommodated instead of cured.
Globally, autism is estimated to affect 24.8 million people . In the 2000s, the number of people affected was estimated at 1โ2 per 1,000 people worldwide. In the developed countries, about 1.5% of children are diagnosed with ASD , from 0.7% in 2000 in the United States. It occurs four-to-five times more often in males than females. The number of people diagnosed has increased dramatically since the 1960s, which may be partly due to changes in diagnostic practice. The question of whether actual rates have increased is unresolved.
Autism is a highly variable, neurodevelopmental disorder whose symptoms first appears during infancy or childhood, and generally follows a steady course without remission. People with autism may be severely impaired in some respects but average, or even superior, in others. Overt symptoms gradually begin after the age of six months, become established by age two or three years and tend to continue through adulthood, although often in more muted form. It is distinguished by a characteristic triad of symptoms: impairments in social interaction, impairments in communication, and repetitive behavior. Other aspects, such as atypical eating, are also common but are not essential for diagnosis. Individual symptoms of autism occur in the general population and appear not to associate highly, without a sharp line separating pathologically severe from common traits.
Social deficits distinguish autism and the related autism spectrum disorders (ASD; see Classification) from other developmental disorders. People with autism have social impairments and often lack the intuition about others that many people take for granted. Noted autistic Temple Grandin described her inability to understand the social communication of neurotypicals, or people with typical neural development, as leaving her feeling "like an anthropologist on Mars".
Unusual social development becomes apparent early in childhood. Autistic infants show less attention to social stimuli, smile and look at others less often, and respond less to their own name. Autistic toddlers differ more strikingly from social norms; for example, they have less eye contact and turn-taking, and do not have the ability to use simple movements to express themselves, such as pointing at things. Three- to five-year-old children with autism are less likely to exhibit social understanding, approach others spontaneously, imitate and respond to emotions, communicate nonverbally, and take turns with others. However, they do form attachments to their primary caregivers. Most children with autism display moderately less attachment security than neurotypical children, although this difference disappears in children with higher mental development or less pronounced autistic traits. Older children and adults with ASD perform worse on tests of face and emotion recognition although this may be partly due to a lower ability to define a person's own emotions.
Children with high-functioning autism have more intense and frequent loneliness compared to non-autistic peers, despite the common belief that children with autism prefer to be alone. Making and maintaining friendships often proves to be difficult for those with autism. For them, the quality of friendships, not the number of friends, predicts how lonely they feel. Functional friendships, such as those resulting in invitations to parties, may affect the quality of life more deeply.
There are many anecdotal reports, but few systematic studies, of aggression and violence in individuals with ASD. The limited data suggest that, in children with intellectual disability, autism is associated with aggression, destruction of property, and meltdowns.
About a third to a half of individuals with autism do not develop enough natural speech to meet their daily communication needs. Differences in communication may be present from the first year of life, and may include delayed onset of babbling, unusual gestures, diminished responsiveness, and vocal patterns that are not synchronized with the caregiver. In the second and third years, children with autism have less frequent and less diverse babbling, consonants, words, and word combinations; their gestures are less often integrated with words. Children with autism are less likely to make requests or share experiences, and are more likely to simply repeat others' words (echolalia) or reverse pronouns. Joint attention seems to be necessary for functional speech, and deficits in joint attention seem to distinguish infants with ASD. For example, they may look at a pointing hand instead of the pointed-at object, and they consistently fail to point at objects in order to comment on or share an experience. Children with autism may have difficulty with imaginative play and with developing symbols into language.
In a pair of studies, high-functioning children with autism aged 8โ15 performed equally well as, and as adults better than, individually matched controls at basic language tasks involving vocabulary and spelling. Both autistic groups performed worse than controls at complex language tasks such as figurative language, comprehension and inference. As people are often sized up initially from their basic language skills, these studies suggest that people speaking to autistic individuals are more likely to overestimate what their audience comprehends.
Autistic individuals can display many forms of repetitive or restricted behavior, which the Repetitive Behavior Scale-Revised (RBS-R) categorizes as follows.
No single repetitive or self-injurious behavior seems to be specific to autism, but autism appears to have an elevated pattern of occurrence and severity of these behaviors.
Autistic individuals may have symptoms that are independent of the diagnosis, but that can affect the individual or the family.
An estimated 0.5% to 10% of individuals with ASD show unusual abilities, ranging from splinter skills such as the memorization of trivia to the extraordinarily rare talents of prodigious autistic savants. Many individuals with ASD show superior skills in perception and attention, relative to the general population. Sensory abnormalities are found in over 90% of those with autism, and are considered core features by some, although there is no good evidence that sensory symptoms differentiate autism from other developmental disorders. Differences are greater for under-responsivity (for example, walking into things) than for over-responsivity (for example, distress from loud noises) or for sensation seeking (for example, rhythmic movements). An estimated 60โ80% of autistic people have motor signs that include poor muscle tone, poor motor planning, and toe walking; deficits in motor coordination are pervasive across ASD and are greater in autism proper. Unusual eating behavior occurs in about three-quarters of children with ASD, to the extent that it was formerly a diagnostic indicator. Selectivity is the most common problem, although eating rituals and food refusal also occur.
There is tentative evidence that autism occurs more frequently in people with gender dysphoria.
Gastrointestinal problems are one of the most commonly associated medical disorders in people with autism. These are linked to greater social impairment, irritability, behavior and sleep problems, language impairments and mood changes.
Parents of children with ASD have higher levels of stress. Siblings of children with ASD report greater admiration of and less conflict with the affected sibling than siblings of unaffected children and were similar to siblings of children with Down syndrome in these aspects of the sibling relationship. However, they reported lower levels of closeness and intimacy than siblings of children with Down syndrome; siblings of individuals with ASD have greater risk of negative well-being and poorer sibling relationships as adults.
It has long been presumed that there is a common cause at the genetic, cognitive, and neural levels for autism's characteristic triad of symptoms. However, there is increasing suspicion that autism is instead a complex disorder whose core aspects have distinct causes that often co-occur.
Autism has a strong genetic basis, although the genetics of autism are complex and it is unclear whether ASD is explained more by rare mutations with major effects, or by rare multigene interactions of common genetic variants. Complexity arises due to interactions among multiple genes, the environment, and epigenetic factors which do not change DNA sequencing but are heritable and influence gene expression. Many genes have been associated with autism through sequencing the genomes of affected individuals and their parents. Studies of twins suggest that heritability is 0.7 for autism and as high as 0.9 for ASD, and siblings of those with autism are about 25 times more likely to be autistic than the general population. However, most of the mutations that increase autism risk have not been identified. Typically, autism cannot be traced to a Mendelian (single-gene) mutation or to a single chromosome abnormality, and none of the genetic syndromes associated with ASDs have been shown to selectively cause ASD. Numerous candidate genes have been located, with only small effects attributable to any particular gene. Most loci individually explain less than 1% of cases of autism. The large number of autistic individuals with unaffected family members may result from spontaneous structural variationโsuch as deletions, duplications or inversions in genetic material during meiosis. Hence, a substantial fraction of autism cases may be traceable to genetic causes that are highly heritable but not inherited: that is, the mutation that causes the autism is not present in the parental genome. Autism may be underdiagnosed in women and girls due to an assumption that it is primarily a male condition, but genetic phenomena such as imprinting and X linkage have the ability to raise the frequency and severity of conditions in males, and theories have been put forward for a genetic reason why males are diagnosed more often, such as the imprinted brain theory and the extreme male brain theory.
Maternal nutrition and inflammation during preconception and pregnancy influences fetal neurodevelopment. Intrauterine growth restriction is associated with ASD, in both term and preterm infants. Maternal inflammatory and autoimmune diseases may damage fetal tissues, aggravating a genetic problem or damaging the nervous system.
Exposure to air pollution during pregnancy, especially heavy metals and particulates, may increase the risk of autism. Environmental factors that have been claimed without evidence to contribute to or exacerbate autism include certain foods, infectious diseases, solvents, PCBs, phthalates and phenols used in plastic products, pesticides, brominated flame retardants, alcohol, smoking, illicit drugs, vaccines, and prenatal stress. Some, such as the MMR vaccine, have been completely disproven.
Parents may first become aware of autistic symptoms in their child around the time of a routine vaccination. This has led to unsupported theories blaming vaccine "overload", a vaccine preservative, or the MMR vaccine for causing autism. The latter theory was supported by a litigation-funded study that has since been shown to have been "an elaborate fraud". Although these theories lack convincing scientific evidence and are biologically implausible, parental concern about a potential vaccine link with autism has led to lower rates of childhood immunizations, outbreaks of previously controlled childhood diseases in some countries, and the preventable deaths of several children.
Autism's symptoms result from maturation-related changes in various systems of the brain. How autism occurs is not well understood. Its mechanism can be divided into two areas: the pathophysiology of brain structures and processes associated with autism, and the neuropsychological linkages between brain structures and behaviors. The behaviors appear to have multiple pathophysiologies.
There is evidence that gutโbrain axis abnormalities may be involved. A 2015 review proposed that immune dysregulation, gastrointestinal inflammation, malfunction of the autonomic nervous system, gut flora alterations, and food metabolites may cause brain neuroinflammation and dysfunction. A 2016 review concludes that enteric nervous system abnormalities might play a role in neurological disorders such as autism. Neural connections and the immune system are a pathway that may allow diseases originated in the intestine to spread to the brain.
Several lines of evidence point to synaptic dysfunction as a cause of autism. Some rare mutations may lead to autism by disrupting some synaptic pathways, such as those involved with cell adhesion. Gene replacement studies in mice suggest that autistic symptoms are closely related to later developmental steps that depend on activity in synapses and on activity-dependent changes. All known teratogens (agents that cause birth defects) related to the risk of autism appear to act during the first eight weeks from conception, and though this does not exclude the possibility that autism can be initiated or affected later, there is strong evidence that autism arises very early in development.
Diagnosis is based on behavior, not cause or mechanism. Under the DSM-5, autism is characterized by persistent deficits in social communication and interaction across multiple contexts, as well as restricted, repetitive patterns of behavior, interests, or activities. These deficits are present in early childhood, typically before age three, and lead to clinically significant functional impairment. Sample symptoms include lack of social or emotional reciprocity, stereotyped and repetitive use of language or idiosyncratic language, and persistent preoccupation with unusual objects. The disturbance must not be better accounted for by Rett syndrome, intellectual disability or global developmental delay. ICD-10 uses essentially the same definition.
Several diagnostic instruments are available. Two are commonly used in autism research: the Autism Diagnostic Interview-Revised (ADI-R) is a semistructured parent interview, and the Autism Diagnostic Observation Schedule (ADOS) uses observation and interaction with the child. The Childhood Autism Rating Scale (CARS) is used widely in clinical environments to assess severity of autism based on observation of children. The Diagnostic interview for social and communication disorders (DISCO) may also be used.
A pediatrician commonly performs a preliminary investigation by taking developmental history and physically examining the child. If warranted, diagnosis and evaluations are conducted with help from ASD specialists, observing and assessing cognitive, communication, family, and other factors using standardized tools, and taking into account any associated medical conditions. A pediatric neuropsychologist is often asked to assess behavior and cognitive skills, both to aid diagnosis and to help recommend educational interventions. A differential diagnosis for ASD at this stage might also consider intellectual disability, hearing impairment, and a specific language impairment such as LandauโKleffner syndrome. The presence of autism can make it harder to diagnose coexisting psychiatric disorders such as depression.
Clinical genetics evaluations are often done once ASD is diagnosed, particularly when other symptoms already suggest a genetic cause. Although genetic technology allows clinical geneticists to link an estimated 40% of cases to genetic causes, consensus guidelines in the US and UK are limited to high-resolution chromosome and fragile X testing. A genotype-first model of diagnosis has been proposed, which would routinely assess the genome's copy number variations. As new genetic tests are developed several ethical, legal, and social issues will emerge. Commercial availability of tests may precede adequate understanding of how to use test results, given the complexity of autism's genetics. Metabolic and neuroimaging tests are sometimes helpful, but are not routine.
ASD can sometimes be diagnosed by age 14 months, although diagnosis becomes increasingly stable over the first three years of life: for example, a one-year-old who meets diagnostic criteria for ASD is less likely than a three-year-old to continue to do so a few years later. In the UK the National Autism Plan for Children recommends at most 30 weeks from first concern to completed diagnosis and assessment, though few cases are handled that quickly in practice. Although the symptoms of autism and ASD begin early in childhood, they are sometimes missed; years later, adults may seek diagnoses to help them or their friends and family understand themselves, to help their employers make adjustments, or in some locations to claim disability living allowances or other benefits. Girls are often diagnosed later than boys.
Underdiagnosis and overdiagnosis are problems in marginal cases, and much of the recent increase in the number of reported ASD cases is likely due to changes in diagnostic practices. The increasing popularity of drug treatment options and the expansion of benefits has given providers incentives to diagnose ASD, resulting in some overdiagnosis of children with uncertain symptoms. Conversely, the cost of screening and diagnosis and the challenge of obtaining payment can inhibit or delay diagnosis. It is particularly hard to diagnose autism among the visually impaired, partly because some of its diagnostic criteria depend on vision, and partly because autistic symptoms overlap with those of common blindness syndromes or blindisms.
Autism is one of the five pervasive developmental disorders (PDD), which are characterized by widespread abnormalities of social interactions and communication, and severely restricted interests and highly repetitive behavior. These symptoms do not imply sickness, fragility, or emotional disturbance.
Of the five PDD forms, Asperger syndrome is closest to autism in signs and likely causes; Rett syndrome and childhood disintegrative disorder share several signs with autism, but may have unrelated causes; PDD not otherwise specified (PDD-NOS; also called "atypical autism") is diagnosed when the criteria are not met for a more specific disorder. Unlike with autism, people with Asperger syndrome have no substantial delay in language development. The terminology of autism can be bewildering, with autism, Asperger syndrome and PDD-NOS often called the "autism spectrum disorders" (ASD) or sometimes the "autistic disorders", whereas autism itself is often called "autistic disorder", "childhood autism", or "infantile autism". In this article, "autism" refers to the classic autistic disorder; in clinical practice, though, "autism", "ASD", and "PDD" are often used interchangeably. ASD, in turn, is a subset of the broader autism phenotype, which describes individuals who may not have ASD but do have autistic-like traits, such as avoiding eye contact.
Autism can also be divided into syndromal and non-syndromal autism; the syndromal autism is associated with severe or profound intellectual disability or a congenital syndrome with physical symptoms, such as tuberous sclerosis. Although individuals with Asperger syndrome tend to perform better cognitively than those with autism, the extent of the overlap between Asperger syndrome, HFA, and non-syndromal autism is unclear.
Some studies have reported diagnoses of autism in children due to a loss of language or social skills, as opposed to a failure to make progress, typically from 15 to 30 months of age. The validity of this distinction remains controversial; it is possible that regressive autism is a specific subtype, or that there is a continuum of behaviors between autism with and without regression.
Research into causes has been hampered by the inability to identify biologically meaningful subgroups within the autistic population and by the traditional boundaries between the disciplines of psychiatry, psychology, neurology and pediatrics. Newer technologies such as fMRI and diffusion tensor imaging can help identify biologically relevant phenotypes (observable traits) that can be viewed on brain scans, to help further neurogenetic studies of autism; one example is lowered activity in the fusiform face area of the brain, which is associated with impaired perception of people versus objects. It has been proposed to classify autism using genetics as well as behavior.
Autism has long been thought to cover a wide spectrum, ranging from individuals with severe impairmentsโwho may be silent, developmentally disabled, and prone to frequent repetitive behavior such as hand flapping and rockingโto high functioning individuals who may have active but distinctly odd social approaches, narrowly focused interests, and verbose, pedantic communication. Because the behavior spectrum is continuous, boundaries between diagnostic categories are necessarily somewhat arbitrary. Sometimes the syndrome is divided into low-, medium- or high-functioning autism (LFA, MFA, and HFA), based on IQ thresholds. Some people have called for an end to the terms "high-functioning" and "low-functioning" due to lack of nuance and the potential for a person's needs or abilities to be overlooked.
About half of parents of children with ASD notice their child's unusual behaviors by age 18 months, and about four-fifths notice by age 24 months. According to an article, failure to meet any of the following milestones "is an absolute indication to proceed with further evaluations. Delay in referral for such testing may delay early diagnosis and treatment and affect the long-term outcome".
The United States Preventive Services Task Force in 2016 found it was unclear if screening was beneficial or harmful among children in whom there is no concerns. The Japanese practice is to screen all children for ASD at 18 and 24 months, using autism-specific formal screening tests. In contrast, in the UK, children whose families or doctors recognize possible signs of autism are screened. It is not known which approach is more effective. Screening tools include the Modified Checklist for Autism in Toddlers (M-CHAT), the Early Screening of Autistic Traits Questionnaire, and the First Year Inventory; initial data on M-CHAT and its predecessor, the Checklist for Autism in Toddlers (CHAT), on children aged 18โ30 months suggests that it is best used in a clinical setting and that it has low sensitivity (many false-negatives) but good specificity (few false-positives). It may be more accurate to precede these tests with a broadband screener that does not distinguish ASD from other developmental disorders. Screening tools designed for one culture's norms for behaviors like eye contact may be inappropriate for a different culture. Although genetic screening for autism is generally still impractical, it can be considered in some cases, such as children with neurological symptoms and dysmorphic features.
While infection with rubella during pregnancy causes fewer than 1% of cases of autism, vaccination against rubella can prevent many of those cases.
The main goals when treating children with autism are to lessen associated deficits and family distress, and to increase quality of life and functional independence. In general, higher IQs are correlated with greater responsiveness to treatment and improved treatment outcomes. No single treatment is best and treatment is typically tailored to the child's needs. Families and the educational system are the main resources for treatment. Services should be carried out by behavior analysts, special education teachers, speech pathologists, and licensed psychologists. Studies of interventions have methodological problems that prevent definitive conclusions about efficacy. However, the development of evidence-based interventions has advanced in recent years. Although many psychosocial interventions have some positive evidence, suggesting that some form of treatment is preferable to no treatment, the methodological quality of systematic reviews of these studies has generally been poor, their clinical results are mostly tentative, and there is little evidence for the relative effectiveness of treatment options. Intensive, sustained special education programs and behavior therapy early in life can help children acquire self-care, communication, and job skills, and often improve functioning and decrease symptom severity and maladaptive behaviors; claims that intervention by around age three years is crucial are not substantiated. While medications have not been found to help with core symptoms, they may be used for associated symptoms, such as irritability, inattention, or repetitive behavior patterns.
Educational interventions often used include applied behavior analysis (ABA), developmental models, structured teaching, speech and language therapy, social skills therapy, and occupational therapy. Among these approaches, interventions either treat autistic features comprehensively, or focalize treatment on a specific area of deficit. The quality of research for early intensive behavioral intervention (EIBI)โa treatment procedure incorporating over thirty hours per week of the structured type of ABA that is carried out with very young childrenโis currently low, and more vigorous research designs with larger sample sizes are needed. Two theoretical frameworks outlined for early childhood intervention include structured and naturalistic ABA interventions, and developmental social pragmatic models (DSP). One interventional strategy utilizes a parent training model, which teaches parents how to implement various ABA and DSP techniques, allowing for parents to disseminate interventions themselves. Various DSP programs have been developed to explicitly deliver intervention systems through at-home parent implementation. Despite the recent development of parent training models, these interventions have demonstrated effectiveness in numerous studies, being evaluated as a probable efficacious mode of treatment.
Early, intensive ABA therapy has demonstrated effectiveness in enhancing communication and adaptive functioning in preschool children; it is also well-established for improving the intellectual performance of that age group. Similarly, a teacher-implemented intervention that utilizes a more naturalistic form of ABA combined with a developmental social pragmatic approach has been found to be beneficial in improving social-communication skills in young children, although there is less evidence in its treatment of global symptoms. Neuropsychological reports are often poorly communicated to educators, resulting in a gap between what a report recommends and what education is provided. It is not known whether treatment programs for children lead to significant improvements after the children grow up, and the limited research on the effectiveness of adult residential programs shows mixed results. The appropriateness of including children with varying severity of autism spectrum disorders in the general education population is a subject of current debate among educators and researchers.
Medications may be used to treat ASD symptoms that interfere with integrating a child into home or school when behavioral treatment fails. They may also be used for associated health problems, such as ADHD or anxiety. More than half of US children diagnosed with ASD are prescribed psychoactive drugs or anticonvulsants, with the most common drug classes being antidepressants, stimulants, and antipsychotics. The atypical antipsychotic drugs risperidone and aripiprazole are FDA-approved for treating associated aggressive and self-injurious behaviors. However, their side effects must be weighed against their potential benefits, and people with autism may respond atypically. Side effects, for example, may include weight gain, tiredness, drooling, and aggression. SSRI antidepressants, such as fluoxetine and fluvoxamine, have been shown to be effective in reducing repetitive and ritualistic behaviors, while the stimulant medication methylphenidate is beneficial for some children with co-morbid inattentiveness or hyperactivity. There is scant reliable research about the effectiveness or safety of drug treatments for adolescents and adults with ASD. No known medication relieves autism's core symptoms of social and communication impairments. Experiments in mice have reversed or reduced some symptoms related to autism by replacing or modulating gene function, suggesting the possibility of targeting therapies to specific rare mutations known to cause autism.
Although many alternative therapies and interventions are available, few are supported by scientific studies. Treatment approaches have little empirical support in quality-of-life contexts, and many programs focus on success measures that lack predictive validity and real-world relevance. Some alternative treatments may place the child at risk. The preference that children with autism have for unconventional foods can lead to reduction in bone cortical thickness with this being greater in those on casein-free diets, as a consequence of the low intake of calcium and vitamin D; however, suboptimal bone development in ASD has also been associated with lack of exercise and gastrointestinal disorders. In 2005, botched chelation therapy killed a five-year-old child with autism. Chelation is not recommended for people with ASD since the associated risks outweigh any potential benefits. Another alternative medicine practice with no evidence is CEASE therapy, a mixture of homeopathy, supplements, and 'vaccine detoxing'.
Although popularly used as an alternative treatment for people with autism, as of 2018 there is no good evidence to recommend a gluten- and casein-free diet as a standard treatment. A 2018 review concluded that it may be a therapeutic option for specific groups of children with autism, such as those with known food intolerances or allergies, or with food intolerance markers. The authors analyzed the prospective trials conducted to date that studied the efficacy of the gluten- and casein-free diet in children with ASD (4 in total). All of them compared gluten- and casein-free diet versus normal diet with a control group (2 double-blind randomized controlled trials, 1 double-blind crossover trial, 1 single-blind trial). In two of the studies, whose duration was 12 and 24 months, a significant improvement in ASD symptoms (efficacy rate 50%) was identified. In the other two studies, whose duration was 3 months, no significant effect was observed. The authors concluded that a longer duration of the diet may be necessary to achieve the improvement of the ASD symptoms. Other problems documented in the trials carried out include transgressions of the diet, small sample size, the heterogeneity of the participants and the possibility of a placebo effect.
In the subset of people who have gluten sensitivity there is limited evidence that suggests that a gluten-free diet may improve some autistic behaviors.
There is tentative evidence that music therapy may improve social interactions, verbal communication, and non-verbal communication skills. There has been early research looking at hyperbaric treatments in children with autism. Studies on pet therapy have shown positive effects.
There is no known cure. The degree of symptoms can decrease, occasionally to the extent that people lose their diagnosis of ASD; this occurs sometimes after intensive treatment and sometimes not. It is not known how often recovery happens; reported rates in unselected samples have ranged from 3% to 25%. Most children with autism acquire language by age five or younger, though a few have developed communication skills in later years. Many children with autism lack social support, future employment opportunities or self-determination. Although core difficulties tend to persist, symptoms often become less severe with age.
Few high-quality studies address long-term prognosis. Some adults show modest improvement in communication skills, but a few decline; no study has focused on autism after midlife. Acquiring language before age six, having an IQ above 50, and having a marketable skill all predict better outcomes; independent living is unlikely with severe autism.
Many individuals with autism face significant obstacles in transitioning to adulthood. Compared to the general population individuals with autism are more likely to be unemployed and to have never had a job. About half of people in their 20s with autism are not employed.
Most recent reviews tend to estimate a prevalence of 1โ2 per 1,000 for autism and close to 6 per 1,000 for ASD as of 2007. A 2016 survey in the United States reported a rate of 25 per 1,000 children for ASD. Globally, autism affects an estimated 24.8 million people , while Asperger syndrome affects a further 37.2 million. In 2012, the NHS estimated that the overall prevalence of autism among adults aged 18 years and over in the UK was 1.1%. Rates of PDD-NOS's has been estimated at 3.7 per 1,000, Asperger syndrome at roughly 0.6 per 1,000, and childhood disintegrative disorder at 0.02 per 1,000. CDC estimates about 1 out of 59 (1.7%) for 2014, an increase from 1 out of every 68 children (1.5%) for 2010.
The number of reported cases of autism increased dramatically in the 1990s and early 2000s. This increase is largely attributable to changes in diagnostic practices, referral patterns, availability of services, age at diagnosis, and public awareness, though unidentified environmental risk factors cannot be ruled out. The available evidence does not rule out the possibility that autism's true prevalence has increased; a real increase would suggest directing more attention and funding toward changing environmental factors instead of continuing to focus on genetics.
Boys are at higher risk for ASD than girls. The sex ratio averages 4.3:1 and is greatly modified by cognitive impairment: it may be close to 2:1 with intellectual disability and more than 5.5:1 without. Several theories about the higher prevalence in males have been investigated, but the cause of the difference is unconfirmed; one theory is that females are underdiagnosed.
Although the evidence does not implicate any single pregnancy-related risk factor as a cause of autism, the risk of autism is associated with advanced age in either parent, and with diabetes, bleeding, and use of psychiatric drugs in the mother during pregnancy. The risk is greater with older fathers than with older mothers; two potential explanations are the known increase in mutation burden in older sperm, and the hypothesis that men marry later if they carry genetic liability and show some signs of autism. Most professionals believe that race, ethnicity, and socioeconomic background do not affect the occurrence of autism.
Several other conditions are common in children with autism. They include:
A few examples of autistic symptoms and treatments were described long before autism was named. The "Table Talk" of Martin Luther, compiled by his notetaker, Mathesius, contains the story of a 12-year-old boy who may have been severely autistic. Luther reportedly thought the boy was a soulless mass of flesh possessed by the devil, and suggested that he be suffocated, although a later critic has cast doubt on the veracity of this report. The earliest well-documented case of autism is that of Hugh Blair of Borgue, as detailed in a 1747 court case in which his brother successfully petitioned to annul Blair's marriage to gain Blair's inheritance. The Wild Boy of Aveyron, a feral child caught in 1798, showed several signs of autism; the medical student Jean Itard treated him with a behavioral program designed to help him form social attachments and to induce speech via imitation.
The New Latin word "autismus" (English translation "autism") was coined by the Swiss psychiatrist Eugen Bleuler in 1910 as he was defining symptoms of schizophrenia. He derived it from the Greek word "autรณs" (ฮฑแฝฯฯฯ, meaning "self"), and used it to mean morbid self-admiration, referring to "autistic withdrawal of the patient to his fantasies, against which any influence from outside becomes an intolerable disturbance". A Soviet child psychiatrist, Grunya Sukhareva, described a similar syndrome that was published in Russian in 1925, and in German in 1926.
The word "autism" first took its modern sense in 1938 when Hans Asperger of the Vienna University Hospital adopted Bleuler's terminology "autistic psychopaths" in a lecture in German about child psychology. Asperger was investigating an ASD now known as Asperger syndrome, though for various reasons it was not widely recognized as a separate diagnosis until 1981. Leo Kanner of the Johns Hopkins Hospital first used "autism" in its modern sense in English when he introduced the label "early infantile autism" in a 1943 report of 11 children with striking behavioral similarities. Almost all the characteristics described in Kanner's first paper on the subject, notably "autistic aloneness" and "insistence on sameness", are still regarded as typical of the autistic spectrum of disorders. It is not known whether Kanner derived the term independently of Asperger.
Donald Triplett was the first person diagnosed with autism. He was diagnosed by Kanner after being first examined in 1938, and was labeled as "case 1". Triplett was noted for his savant abilities, particularly being able to name musical notes played on a piano and to mentally multiply numbers. His father, Oliver, described him as socially withdrawn but interested in number patterns, music notes, letters of the alphabet, and U.S. president pictures. By the age of 2, he had the ability to recite the 23rd Psalm and memorized 25 questions and answers from the Presbyterian catechism. He was also interested in creating musical chords.
Kanner's reuse of "autism" led to decades of confused terminology like "infantile schizophrenia", and child psychiatry's focus on maternal deprivation led to misconceptions of autism as an infant's response to "refrigerator mothers". Starting in the late 1960s autism was established as a separate syndrome.
As late as the mid-1970s there was little evidence of a genetic role in autism; while in 2007 it was believed to be one of the most heritable psychiatric conditions. Although the rise of parent organizations and the destigmatization of childhood ASD have affected how ASD is viewed, parents continue to feel social stigma in situations where their child's autistic behavior is perceived negatively, and many primary care physicians and medical specialists express some beliefs consistent with outdated autism research.
It took until 1980 for the DSM-III to differentiate autism from childhood schizophrenia. In 1987, the DSM-III-R provided a checklist for diagnosing autism. In May 2013, the DSM-5 was released, updating the classification for pervasive developmental disorders. The grouping of disorders, including PDD-NOS, autism, Asperger syndrome, Rett syndrome, and CDD, has been removed and replaced with the general term of Autism Spectrum Disorders. The two categories that exist are impaired social communication and/or interaction, and restricted and/or repetitive behaviors.
The Internet has helped autistic individuals bypass nonverbal cues and emotional sharing that they find difficult to deal with, and has given them a way to form online communities and work remotely. Societal and cultural aspects of autism have developed: some in the community seek a cure, while others believe that autism is simply another way of being.
An autistic culture has emerged, accompanied by the autistic rights and neurodiversity movements. Events include World Autism Awareness Day, Autism Sunday, Autistic Pride Day, Autreat, and others. Organizations dedicated to promoting awareness of autism include Autistic Self Advocacy Network, Aspies For Freedom, Autism National Committee, and Autism Society of America. At the same time, some organizations, including Autism Speaks, have been condemned by disability rights organizations for failing to support autistic people. Social-science scholars study those with autism in hopes to learn more about "autism as a culture, transcultural comparisons... and research on social movements." While most autistic individuals do not have savant skills, many have been successful in their fields.
The autism rights movement is a social movement within the context of disability rights that emphasizes the concept of neurodiversity, viewing the autism spectrum as a result of natural variations in the human brain rather than a disorder to be cured. The autism rights movement advocates for including greater acceptance of autistic behaviors; therapies that focus on coping skills rather than on imitating the behaviors of those without autism, and the recognition of the autistic community as a minority group. Autism rights or neurodiversity advocates believe that the autism spectrum is genetic and should be accepted as a natural expression of the human genome. This perspective is distinct from two other likewise distinct views: the medical perspective, that autism is caused by a genetic defect and should be addressed by targeting the autism gene(s), and fringe theories that autism is caused by environmental factors such as vaccines. A common criticism against autistic activists is that the majority of them are "high-functioning" or have Asperger syndrome and do not represent the views of "low-functioning" autistic people.
About half of autistics are unemployed, and one third of those with graduate degrees may be unemployed. Among autistics who find work, most are employed in sheltered settings working for wages below the national minimum. While employers state hiring concerns about productivity and supervision, experienced employers of autistics give positive reports of above average memory and detail orientation as well as a high regard for rules and procedure in autistic employees. A majority of the economic burden of autism is caused by decreased earnings in the job market. Some studies also find decreased earning among parents who care for autistic children.
</doc>
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/SQUAD/sample.json
|
{
"version": 2.0,
"data": [
{
"id": "56ddde6b9a695914005b9628",
"question": "In what country is Normandy located?",
"context": "The Normans (Norman: Nourmands; French: Normands; Latin: Normanni) were the people who in the 10th and 11th centuries gave their name to Normandy, a region in France. They were descended from Norse (\"Norman\" comes from \"Norseman\") raiders and pirates from Denmark, Iceland and Norway who, under their leader Rollo, agreed to swear fealty to King Charles III of West Francia. Through generations of assimilation and mixing with the native Frankish and Roman-Gaulish populations, their descendants would gradually merge with the Carolingian-based cultures of West Francia. The distinct cultural and ethnic identity of the Normans emerged initially in the first half of the 10th century, and it continued to evolve over the succeeding centuries.",
"answers": {
"answer_start": [
159,
159,
159,
159
],
"text": [
"France",
"France",
"France",
"France"
]
}
},
{
"id": "56ddde6b9a695914005b9629",
"question": "When were the Normans in Normandy?",
"context": "The Normans (Norman: Nourmands; French: Normands; Latin: Normanni) were the people who in the 10th and 11th centuries gave their name to Normandy, a region in France. They were descended from Norse (\"Norman\" comes from \"Norseman\") raiders and pirates from Denmark, Iceland and Norway who, under their leader Rollo, agreed to swear fealty to King Charles III of West Francia. Through generations of assimilation and mixing with the native Frankish and Roman-Gaulish populations, their descendants would gradually merge with the Carolingian-based cultures of West Francia. The distinct cultural and ethnic identity of the Normans emerged initially in the first half of the 10th century, and it continued to evolve over the succeeding centuries.",
"answers": {
"answer_start": [
94,
87,
94,
94
],
"text": [
"10th and 11th centuries",
"in the 10th and 11th centuries",
"10th and 11th centuries",
"10th and 11th centuries"
]
}
},
{
"id": "56ddde6b9a695914005b962a",
"question": "From which countries did the Norse originate?",
"context": "The Normans (Norman: Nourmands; French: Normands; Latin: Normanni) were the people who in the 10th and 11th centuries gave their name to Normandy, a region in France. They were descended from Norse (\"Norman\" comes from \"Norseman\") raiders and pirates from Denmark, Iceland and Norway who, under their leader Rollo, agreed to swear fealty to King Charles III of West Francia. Through generations of assimilation and mixing with the native Frankish and Roman-Gaulish populations, their descendants would gradually merge with the Carolingian-based cultures of West Francia. The distinct cultural and ethnic identity of the Normans emerged initially in the first half of the 10th century, and it continued to evolve over the succeeding centuries.",
"answers": {
"answer_start": [
256,
256,
256,
256
],
"text": [
"Denmark, Iceland and Norway",
"Denmark, Iceland and Norway",
"Denmark, Iceland and Norway",
"Denmark, Iceland and Norway"
]
}
},
{
"id": "5ad39d53604f3c001a3fe8d3",
"question": "Who did King Charles III swear fealty to?",
"context": "The Normans (Norman: Nourmands; French: Normands; Latin: Normanni) were the people who in the 10th and 11th centuries gave their name to Normandy, a region in France. They were descended from Norse (\"Norman\" comes from \"Norseman\") raiders and pirates from Denmark, Iceland and Norway who, under their leader Rollo, agreed to swear fealty to King Charles III of West Francia. Through generations of assimilation and mixing with the native Frankish and Roman-Gaulish populations, their descendants would gradually merge with the Carolingian-based cultures of West Francia. The distinct cultural and ethnic identity of the Normans emerged initially in the first half of the 10th century, and it continued to evolve over the succeeding centuries.",
"answers": {
"answer_start": [],
"text": []
}
},
{
"id": "5ad39d53604f3c001a3fe8d4",
"question": "When did the Frankish identity emerge?",
"context": "The Normans (Norman: Nourmands; French: Normands; Latin: Normanni) were the people who in the 10th and 11th centuries gave their name to Normandy, a region in France. They were descended from Norse (\"Norman\" comes from \"Norseman\") raiders and pirates from Denmark, Iceland and Norway who, under their leader Rollo, agreed to swear fealty to King Charles III of West Francia. Through generations of assimilation and mixing with the native Frankish and Roman-Gaulish populations, their descendants would gradually merge with the Carolingian-based cultures of West Francia. The distinct cultural and ethnic identity of the Normans emerged initially in the first half of the 10th century, and it continued to evolve over the succeeding centuries.",
"answers": {
"answer_start": [],
"text": []
}
},
{
"id": "56dddf4066d3e219004dad5f",
"question": "Who was the duke in the battle of Hastings?",
"context": "The Norman dynasty had a major political, cultural and military impact on medieval Europe and even the Near East. The Normans were famed for their martial spirit and eventually for their Christian piety, becoming exponents of the Catholic orthodoxy into which they assimilated. They adopted the Gallo-Romance language of the Frankish land they settled, their dialect becoming known as Norman, Normaund or Norman French, an important literary language. The Duchy of Normandy, which they formed by treaty with the French crown, was a great fief of medieval France, and under Richard I of Normandy was forged into a cohesive and formidable principality in feudal tenure. The Normans are noted both for their culture, such as their unique Romanesque architecture and musical traditions, and for their significant military accomplishments and innovations. Norman adventurers founded the Kingdom of Sicily under Roger II after conquering southern Italy on the Saracens and Byzantines, and an expedition on behalf of their duke, William the Conqueror, led to the Norman conquest of England at the Battle of Hastings in 1066. Norman cultural and military influence spread from these new European centres to the Crusader states of the Near East, where their prince Bohemond I founded the Principality of Antioch in the Levant, to Scotland and Wales in Great Britain, to Ireland, and to the coasts of north Africa and the Canary Islands.",
"answers": {
"answer_start": [
1022,
1022,
1022
],
"text": [
"William the Conqueror",
"William the Conqueror",
"William the Conqueror"
]
}
},
{
"id": "5ad3a266604f3c001a3fea2b",
"question": "What principality did William the conquerer found?",
"context": "The Norman dynasty had a major political, cultural and military impact on medieval Europe and even the Near East. The Normans were famed for their martial spirit and eventually for their Christian piety, becoming exponents of the Catholic orthodoxy into which they assimilated. They adopted the Gallo-Romance language of the Frankish land they settled, their dialect becoming known as Norman, Normaund or Norman French, an important literary language. The Duchy of Normandy, which they formed by treaty with the French crown, was a great fief of medieval France, and under Richard I of Normandy was forged into a cohesive and formidable principality in feudal tenure. The Normans are noted both for their culture, such as their unique Romanesque architecture and musical traditions, and for their significant military accomplishments and innovations. Norman adventurers founded the Kingdom of Sicily under Roger II after conquering southern Italy on the Saracens and Byzantines, and an expedition on behalf of their duke, William the Conqueror, led to the Norman conquest of England at the Battle of Hastings in 1066. Norman cultural and military influence spread from these new European centres to the Crusader states of the Near East, where their prince Bohemond I founded the Principality of Antioch in the Levant, to Scotland and Wales in Great Britain, to Ireland, and to the coasts of north Africa and the Canary Islands.",
"answers": {
"answer_start": [],
"text": []
}
},
{
"id": "56e16182e3433e1400422e28",
"question": "What branch of theoretical computer science deals with broadly classifying computational problems by difficulty and class of relationship?",
"context": "Computational complexity theory is a branch of the theory of computation in theoretical computer science that focuses on classifying computational problems according to their inherent difficulty, and relating those classes to each other. A computational problem is understood to be a task that is in principle amenable to being solved by a computer, which is equivalent to stating that the problem may be solved by mechanical application of mathematical steps, such as an algorithm.",
"answers": {
"answer_start": [
0,
0,
0
],
"text": [
"Computational complexity theory",
"Computational complexity theory",
"Computational complexity theory"
]
}
},
{
"id": "5ad5316b5b96ef001a10ab76",
"question": "What is a manual application of mathematical steps?",
"context": "Computational complexity theory is a branch of the theory of computation in theoretical computer science that focuses on classifying computational problems according to their inherent difficulty, and relating those classes to each other. A computational problem is understood to be a task that is in principle amenable to being solved by a computer, which is equivalent to stating that the problem may be solved by mechanical application of mathematical steps, such as an algorithm.",
"answers": {
"answer_start": [],
"text": []
}
},
{
"id": "56e16839cd28a01900c67887",
"question": "What measure of a computational problem broadly defines the inherent difficulty of the solution?",
"context": "A problem is regarded as inherently difficult if its solution requires significant resources, whatever the algorithm used. The theory formalizes this intuition, by introducing mathematical models of computation to study these problems and quantifying the amount of resources needed to solve them, such as time and storage. Other complexity measures are also used, such as the amount of communication (used in communication complexity), the number of gates in a circuit (used in circuit complexity) and the number of processors (used in parallel computing). One of the roles of computational complexity theory is to determine the practical limits on what computers can and cannot do.",
"answers": {
"answer_start": [
46,
49,
46
],
"text": [
"if its solution requires significant resources",
"its solution requires significant resources",
"if its solution requires significant resources"
]
}
},
{
"id": "56e16839cd28a01900c67888",
"question": "What method is used to intuitively assess or quantify the amount of resources required to solve a computational problem?",
"context": "A problem is regarded as inherently difficult if its solution requires significant resources, whatever the algorithm used. The theory formalizes this intuition, by introducing mathematical models of computation to study these problems and quantifying the amount of resources needed to solve them, such as time and storage. Other complexity measures are also used, such as the amount of communication (used in communication complexity), the number of gates in a circuit (used in circuit complexity) and the number of processors (used in parallel computing). One of the roles of computational complexity theory is to determine the practical limits on what computers can and cannot do.",
"answers": {
"answer_start": [
176,
176,
176
],
"text": [
"mathematical models of computation",
"mathematical models of computation",
"mathematical models of computation"
]
}
},
{
"id": "56e16839cd28a01900c67889",
"question": "What are two basic primary resources used to guage complexity?",
"context": "A problem is regarded as inherently difficult if its solution requires significant resources, whatever the algorithm used. The theory formalizes this intuition, by introducing mathematical models of computation to study these problems and quantifying the amount of resources needed to solve them, such as time and storage. Other complexity measures are also used, such as the amount of communication (used in communication complexity), the number of gates in a circuit (used in circuit complexity) and the number of processors (used in parallel computing). One of the roles of computational complexity theory is to determine the practical limits on what computers can and cannot do.",
"answers": {
"answer_start": [
305,
305,
305
],
"text": [
"time and storage",
"time and storage",
"time and storage"
]
}
},
{
"id": "5ad532575b96ef001a10ab7f",
"question": "What unit is measured to determine circuit simplicity?",
"context": "A problem is regarded as inherently difficult if its solution requires significant resources, whatever the algorithm used. The theory formalizes this intuition, by introducing mathematical models of computation to study these problems and quantifying the amount of resources needed to solve them, such as time and storage. Other complexity measures are also used, such as the amount of communication (used in communication complexity), the number of gates in a circuit (used in circuit complexity) and the number of processors (used in parallel computing). One of the roles of computational complexity theory is to determine the practical limits on what computers can and cannot do.",
"answers": {
"answer_start": [],
"text": []
}
},
{
"id": "5ad532575b96ef001a10ab80",
"question": "What number is used in perpendicular computing?",
"context": "A problem is regarded as inherently difficult if its solution requires significant resources, whatever the algorithm used. The theory formalizes this intuition, by introducing mathematical models of computation to study these problems and quantifying the amount of resources needed to solve them, such as time and storage. Other complexity measures are also used, such as the amount of communication (used in communication complexity), the number of gates in a circuit (used in circuit complexity) and the number of processors (used in parallel computing). One of the roles of computational complexity theory is to determine the practical limits on what computers can and cannot do.",
"answers": {
"answer_start": [],
"text": []
}
}
]
}
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/swag/sample.json
|
{"ending0": "passes by walking down the street playing their instruments.", "ending1": "has heard approaching them.", "ending2": "arrives and they're outside dancing and asleep.", "ending3": "turns the lead singer watches the performance.", "label": 0, "sent1": "Members of the procession walk down the street holding small horn brass instruments.", "sent2": "A drum line"}
{"ending0": "are playing ping pong and celebrating one left each in quick.", "ending1": "wait slowly towards the cadets.", "ending2": "continues to play as well along the crowd along with the band being interviewed.", "ending3": "continue to play marching, interspersed.", "label": 3, "sent1": "A drum line passes by walking down the street playing their instruments.", "sent2": "Members of the procession"}
{"ending0": "pay the other coaches to cheer as people this chatter dips in lawn sheets.", "ending1": "walk down the street holding small horn brass instruments.", "ending2": "is seen in the background.", "ending3": "are talking a couple of people playing a game of tug of war.", "label": 1, "sent1": "A group of members in green uniforms walks waving flags.", "sent2": "Members of the procession"}
{"ending0": "are playing ping pong and celebrating one left each in quick.", "ending1": "wait slowly towards the cadets.", "ending2": "makes a square call and ends by jumping down into snowy streets where fans begin to take their positions.", "ending3": "play and go back and forth hitting the drums while the audience claps for them.", "label": 3, "sent1": "A drum line passes by walking down the street playing their instruments.", "sent2": "Members of the procession"}
{"ending0": "finishes the song and lowers the instrument.", "ending1": "hits the saxophone and demonstrates how to properly use the racquet.", "ending2": "finishes massage the instrument again and continues.", "ending3": "continues dancing while the man gore the music outside while drums.", "label": 0, "sent1": "The person plays a song on the violin.", "sent2": "The man"}
{"ending0": "finishes playing then marches their tenderly.", "ending1": "walks in frame and rubs on his hands, and then walks into a room.", "ending2": "continues playing guitar while moving from the camera.", "ending3": "plays a song on the violin.", "label": 3, "sent1": "The person holds up the violin to his chin and gets ready.", "sent2": "The person"}
{"ending0": "examines the instrument in his hand.", "ending1": "stops playing the drums and waves over the other boys.", "ending2": "lights the cigarette and sticks his head in.", "ending3": "drags off the vacuum.", "label": 0, "sent1": "A person retrieves an instrument from a closet.", "sent2": "The man"}
{"ending0": "studies a picture of the man playing the violin.", "ending1": "holds up the violin to his chin and gets ready.", "ending2": "stops to speak to the camera again.", "ending3": "puts his arm around the man and backs away.", "label": 1, "sent1": "The man examines the instrument in his hand.", "sent2": "The person"}
{"ending0": "hands her another phone.", "ending1": "takes the drink, then holds it.", "ending2": "looks off then looks at someone.", "ending3": "stares blearily down at the floor.", "label": 3, "sent1": "Someone walks over to the radio.", "sent2": "Someone"}
{"ending0": "looks off then looks at someone.", "ending1": "hands her another phone.", "ending2": "takes the drink, then holds it.", "ending3": "turns on a monitor.", "label": 3, "sent1": "Someone walks over to the radio.", "sent2": "Someone"}
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/COCO/coco_panoptic_annotations.txt
|
[{"id": 8222595, "category_id": 17, "iscrowd": 0, "bbox": [18, 54, 301, 415], "area": 53306}, {"id": 8225432, "category_id": 17, "iscrowd": 0, "bbox": [349, 26, 291, 343], "area": 59627}, {"id": 8798150, "category_id": 63, "iscrowd": 0, "bbox": [1, 0, 639, 474], "area": 174579}, {"id": 14466198, "category_id": 75, "iscrowd": 0, "bbox": [42, 74, 133, 45], "area": 4068}, {"id": 12821912, "category_id": 75, "iscrowd": 0, "bbox": [333, 80, 38, 106], "area": 2118}, {"id": 10898909, "category_id": 93, "iscrowd": 0, "bbox": [0, 0, 640, 480], "area": 2750}]
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/COCO/coco_annotations.txt
|
[{"segmentation": [[333.96, 175.14, 338.26, 134.33, 342.55, 95.67, 348.99, 79.57, 368.32, 80.64, 371.54, 91.38, 364.03, 106.41, 356.51, 145.07, 351.14, 166.55, 350.07, 184.8, 345.77, 185.88, 332.89, 178.36, 332.89, 172.99]], "area": 2120.991099999999, "iscrowd": 0, "image_id": 39769, "bbox": [332.89, 79.57, 38.65, 106.31], "category_id": 75, "id": 1108446}, {"segmentation": [[44.03, 86.01, 112.75, 74.2, 173.96, 77.42, 175.03, 89.23, 170.74, 98.9, 147.11, 102.12, 54.77, 119.3, 53.69, 119.3, 44.03, 113.93, 41.88, 94.6, 41.88, 94.6]], "area": 4052.607, "iscrowd": 0, "image_id": 39769, "bbox": [41.88, 74.2, 133.15, 45.1], "category_id": 75, "id": 1110067}, {"segmentation": [[1.08, 473.53, 633.17, 473.53, 557.66, 376.45, 535.01, 366.74, 489.71, 305.26, 470.29, 318.2, 456.27, 351.64, 413.12, 363.51, 376.45, 358.11, 348.4, 350.56, 363.51, 331.15, 357.03, 288.0, 353.8, 257.8, 344.09, 190.92, 333.3, 177.98, 345.17, 79.82, 284.76, 130.52, 265.35, 151.01, 308.49, 189.84, 317.12, 215.73, 293.39, 243.78, 269.66, 212.49, 235.15, 199.55, 214.65, 193.08, 187.69, 217.89, 159.64, 278.29, 135.91, 313.89, 169.35, 292.31, 203.87, 281.53, 220.04, 292.31, 220.04, 307.42, 175.82, 345.17, 155.33, 360.27, 105.71, 363.51, 85.21, 374.29, 74.43, 366.74, 70.11, 465.98, 42.07, 471.37, 33.44, 457.35, 34.52, 414.2, 29.12, 368.9, 9.71, 291.24, 46.38, 209.26, 99.24, 128.36, 131.6, 107.87, 50.7, 117.57, 40.99, 103.55, 40.99, 85.21, 60.4, 77.66, 141.3, 70.11, 173.66, 72.27, 174.74, 92.76, 204.94, 72.27, 225.44, 62.56, 262.11, 56.09, 292.31, 53.93, 282.61, 81.98, 298.79, 96.0, 310.65, 102.47, 348.4, 74.43, 373.21, 81.98, 430.38, 35.6, 484.31, 23.73, 540.4, 46.38, 593.26, 66.88, 638.56, 80.9, 632.09, 145.62, 581.39, 118.65, 543.64, 130.52, 533.93, 167.19, 512.36, 197.39, 498.34, 218.97, 529.62, 253.48, 549.03, 273.98, 584.63, 276.13, 587.87, 293.39, 566.29, 305.26, 531.78, 298.79, 549.03, 319.28, 576.0, 358.11, 560.9, 376.45, 639.64, 471.37, 639.64, 2.16, 1.08, 0.0]], "area": 176277.55269999994, "iscrowd": 0, "image_id": 39769, "bbox": [1.08, 0.0, 638.56, 473.53], "category_id": 63, "id": 1605237}, {"segmentation": [[1.07, 1.18, 640.0, 3.33, 638.93, 472.59, 4.3, 479.03]], "area": 301552.6694999999, "iscrowd": 0, "image_id": 39769, "bbox": [1.07, 1.18, 638.93, 477.85], "category_id": 65, "id": 1612051}, {"segmentation": [[138.75, 319.38, 148.75, 294.38, 165.0, 246.87, 197.5, 205.63, 247.5, 203.13, 268.75, 216.88, 280.0, 239.38, 293.75, 244.38, 303.75, 241.88, 307.5, 228.13, 318.75, 220.63, 315.0, 200.63, 291.25, 171.88, 265.0, 156.88, 258.75, 148.13, 262.5, 135.63, 282.5, 123.13, 292.5, 115.63, 311.25, 108.13, 313.75, 106.88, 296.25, 93.13, 282.5, 84.38, 292.5, 64.38, 288.75, 60.63, 266.25, 54.38, 232.5, 63.12, 206.25, 70.63, 170.0, 100.63, 136.25, 114.38, 101.25, 138.13, 56.25, 194.38, 27.5, 259.38, 17.5, 299.38, 32.5, 378.13, 31.25, 448.13, 41.25, 469.38, 66.25, 466.88, 70.0, 419.38, 71.25, 391.88, 77.5, 365.63, 113.75, 364.38, 145.0, 360.63, 168.75, 349.38, 191.25, 330.63, 212.5, 319.38, 223.75, 305.63, 206.25, 286.88, 172.5, 288.13]], "area": 53301.618749999994, "iscrowd": 0, "image_id": 39769, "bbox": [17.5, 54.38, 301.25, 415.0], "category_id": 17, "id": 2190839}, {"segmentation": [[543.75, 136.88, 570.0, 114.38, 591.25, 123.13, 616.25, 140.63, 640.0, 143.13, 636.25, 124.37, 605.0, 103.13, 640.0, 103.13, 633.75, 86.88, 587.5, 73.13, 548.75, 49.38, 505.0, 35.63, 462.5, 25.63, 405.0, 48.13, 362.5, 111.88, 347.5, 179.38, 355.0, 220.63, 356.25, 230.63, 365.0, 264.38, 358.75, 266.88, 358.75, 270.63, 356.25, 291.88, 356.25, 325.63, 355.0, 338.13, 350.0, 348.13, 365.0, 354.38, 396.25, 351.88, 423.75, 355.63, 446.25, 350.63, 460.0, 345.63, 462.5, 321.88, 468.75, 306.88, 481.25, 299.38, 516.25, 341.88, 536.25, 368.13, 570.0, 369.38, 578.75, 359.38, 555.0, 330.63, 532.5, 298.13, 563.75, 299.38, 582.5, 298.13, 586.25, 286.88, 578.75, 278.13, 548.75, 269.38, 525.0, 256.88, 505.0, 206.88, 536.25, 161.88, 540.0, 149.38]], "area": 59700.95625, "iscrowd": 0, "image_id": 39769, "bbox": [347.5, 25.63, 292.5, 343.75], "category_id": 17, "id": 2190842}]
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/STS-B/dev.tsv
|
index genre filename year old_index source1 source2 sentence1 sentence2 score
0 main-captions MSRvid 2012test 0000 none none A man with a hard hat is dancing. A man wearing a hard hat is dancing. 5.000
1 main-captions MSRvid 2012test 0002 none none A young child is riding a horse. A child is riding a horse. 4.750
2 main-captions MSRvid 2012test 0003 none none A man is feeding a mouse to a snake. The man is feeding a mouse to the snake. 5.000
3 main-captions MSRvid 2012test 0007 none none A woman is playing the guitar. A man is playing guitar. 2.400
4 main-captions MSRvid 2012test 0008 none none A woman is playing the flute. A man is playing a flute. 2.750
5 main-captions MSRvid 2012test 0010 none none A woman is cutting an onion. A man is cutting onions. 2.615
6 main-captions MSRvid 2012test 0015 none none A man is erasing a chalk board. The man is erasing the chalk board. 5.000
7 main-captions MSRvid 2012test 0023 none none A woman is carrying a boy. A woman is carrying her baby. 2.333
8 main-captions MSRvid 2012test 0027 none none Three men are playing guitars. Three men are on stage playing guitars. 3.750
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/STS-B/train.tsv
|
index genre filename year old_index source1 source2 sentence1 sentence2 score
0 main-captions MSRvid 2012test 0001 none none A plane is taking off. An air plane is taking off. 5.000
1 main-captions MSRvid 2012test 0004 none none A man is playing a large flute. A man is playing a flute. 3.800
2 main-captions MSRvid 2012test 0005 none none A man is spreading shreded cheese on a pizza. A man is spreading shredded cheese on an uncooked pizza. 3.800
3 main-captions MSRvid 2012test 0006 none none Three men are playing chess. Two men are playing chess. 2.600
4 main-captions MSRvid 2012test 0009 none none A man is playing the cello. A man seated is playing the cello. 4.250
5 main-captions MSRvid 2012test 0011 none none Some men are fighting. Two men are fighting. 4.250
6 main-captions MSRvid 2012test 0012 none none A man is smoking. A man is skating. 0.500
7 main-captions MSRvid 2012test 0013 none none The man is playing the piano. The man is playing the guitar. 1.600
8 main-captions MSRvid 2012test 0014 none none A man is playing on a guitar and singing. A woman is playing an acoustic guitar and singing. 2.200
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/GermEval/dev.txt
|
Gleich O
darauf O
entwirft O
er O
seine O
Selbstdarstellung O
" O
Ecce B-OTH
homo I-OTH
" O
in O
enger O
Auseinandersetzung O
mit O
diesem O
Bild O
Jesu B-PER
. O
1980 O
kam O
der O
Crown B-OTH
als O
Versuch O
von O
Toyota B-ORG
, O
sich O
in O
der O
Oberen O
Mittelklasse O
zu O
etablieren O
, O
auch O
nach O
Deutschland B-LOC
. O
โ O
4:26 O
# O
Sometime B-OTH
Ago/La I-OTH
Fiesta I-OTH
โ O
23:18 O
Alle O
Stรผcke O
wurden O
von O
Corea B-PER
komponiert O
mit O
Ausnahme O
der O
einleitenden O
Improvisation O
zu O
Sometime B-OTH
Ago I-OTH
. O
Bis O
2013 O
steigen O
die O
Mittel O
aus O
dem O
EU-Budget B-ORGpart
auf O
rund O
120 O
Millionen O
Euro B-OTH
. O
Daraus O
entwickelte O
sich O
im O
Rokoko B-OTH
die O
Sitte O
des O
gemeinsamen O
Weinens O
im O
Theater O
, O
das O
die O
Standesgrenzen O
innerhalb O
des O
Publikums O
รผberbrรผcken O
sollte O
. O
Die O
Spinne O
hatte O
sie O
mit O
Seidenfรคden O
an O
ihrem O
Schwanz O
gefesselt O
und O
nach O
oben O
gezogen O
. O
In O
Deutschland B-LOC
ist O
nach O
StGB O
eine O
Anwerbung O
fรผr O
die O
Fremdenlegion O
strafbar O
. O
Am O
Donnerstag O
wird O
sich O
zeigen O
, O
ob O
die O
Idee O
der O
DLR-Forscher B-ORGpart
funktioniert O
. O
Der O
sechste O
Lauf O
der O
ADAC B-ORG
GT I-ORG
Mastersstand O
ganz O
klar O
im O
Mittelpunkt O
des O
Motorsport-Wochenendes O
auf O
dem O
Eurospeedway B-ORG
Lausitz I-ORG
. O
Nach O
den O
schwรคcheren O
Vorgaben O
der O
Wall B-ORG
Street I-ORG
vom O
Vortag O
setzten O
die O
deutschen B-LOCderiv
Standardwerte O
ihren O
Konsolidierungskurs O
fort O
. O
Kolb B-PER
war O
seit O
1986 O
im O
Turnverein O
als O
Leiter O
tรคtig O
, O
darunter O
elf O
Jahre O
als O
Hauptleiter O
in O
der O
Mรคnnerriege O
. O
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/GermEval/train.txt
|
Schartau B-PER
sagte O
dem O
" O
Tagesspiegel B-ORG
" O
vom O
Freitag O
, O
Fischer B-PER
sei O
" O
in O
einer O
Weise O
aufgetreten O
, O
die O
alles O
andere O
als O
รผberzeugend O
war O
" O
. O
Firmengrรผnder O
Wolf B-PER
Peter I-PER
Bree I-PER
arbeitete O
Anfang O
der O
siebziger O
Jahre O
als O
Mรถbelvertreter O
, O
als O
er O
einen O
fliegenden O
Hรคndler O
aus O
dem O
Libanon B-LOC
traf O
. O
Ob O
sie O
dabei O
nach O
dem O
Runden O
Tisch O
am O
23. O
April O
in O
Berlin B-LOC
durch O
ein O
pรคdagogisches O
Konzept O
unterstรผtzt O
wird O
, O
ist O
allerdings O
zu O
bezweifeln O
. O
Bayern B-ORG
Mรผnchen I-ORG
ist O
wieder O
alleiniger O
Top- O
Favorit O
auf O
den O
Gewinn O
der O
deutschen B-LOCderiv
Fuรball-Meisterschaft O
. O
Dabei O
hรคtte O
der O
tapfere O
Schlussmann O
allen O
Grund O
gehabt O
, O
sich O
viel O
frรผher O
aufzuregen O
. O
ARD-Programmchef B-ORGpart
Gรผnter B-PER
Struve I-PER
war O
wegen O
eines O
vierwรถchigen O
Urlaubs O
fรผr O
eine O
Stellungnahme O
nicht O
erreichbar O
. O
Alternativ O
sollten O
sich O
die O
Restaurantbetreiber O
aus O
Sicht O
der O
Solingerin B-LOCderiv
zu O
lรคngeren O
รffnungszeiten O
verpflichten O
, O
um O
wartende O
Kunden O
aufzunehmen O
. O
Die O
Deutsche B-ORG
Flugsicherung I-ORG
( O
DFS B-ORG
) O
beschloss O
ein O
Flugverbot O
fรผr O
alle O
internationalen O
Flughรคfen O
mit O
Ausnahme O
der O
beiden O
Berliner B-LOCderiv
Flughรคfen O
bis O
2.00 O
Uhr O
nachts O
. O
New O
Small O
Family O
mit O
E-Motor O
: O
Studie O
E-Up O
! O
Eine O
Schwachstelle O
war O
beispielsweise O
der O
Spiegelkasten O
. O
Denn O
durch O
den O
Einsatz O
moderner O
Fahrzeugtechnik O
( O
Dieseltriebwagen O
) O
und O
schalldรคmmender O
Fenster O
entsteht O
keine O
Einschrรคnkung O
der O
Wohnqualitรคt O
. O
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/GermEval/labels.txt
|
B-LOC
B-LOCderiv
B-LOCpart
B-ORG
B-ORGderiv
B-ORGpart
B-OTH
B-OTHderiv
B-OTHpart
B-PER
B-PERderiv
B-PERpart
I-LOC
I-LOCderiv
I-LOCpart
I-ORG
I-ORGderiv
I-ORGpart
I-OTH
I-OTHderiv
I-OTHpart
I-PER
I-PERderiv
I-PERpart
O
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/wmt_en_ro/val.json
|
{ "translation": { "en": "Brazil's Former Presidential Chief-of-Staff to Stand Trial A federal judge on Tuesday accepted the charges filed against Brazil's former presidential chief of staff for his alleged involvement in a massive corruption scheme at state-owned oil company Petrobras. The federal prosecutor's office said Jose Dirceu will face trial on the corruption, racketeering and money laundering charges filed earlier this month. Fourteen other people will also be tried, including Joao Vaccari Neto, the former treasurer of Brazil's governing Workers' Party and Renato de Souza Duque, Petrobras' former head of corporate services.", "ro": "Fostul ศef al cabinetului prezidenศial brazilian este adus รฎn faศa instanศei Marศi, un judecฤtor federal a acceptat acuzaศiile aduse รฎmpotriva fostului ศef al cabinetului prezidenศial brazilian pentru presupusa implicare a acestuia รฎntr-o schemฤ masivฤ de corupศie privind compania petrolierฤ de stat Petrobras. Biroul procurorului federal a declarat cฤ Jose Dirceu va fi trimis รฎn judecatฤ pentru acuzaศiile de corupศie, รฎnศelฤtorie ศi spฤlare de bani aduse รฎn aceastฤ lunฤ. Alte paisprezece persoane vor fi judecate, printre acestea numฤrรขndu-se Joao Vaccari Neto, fostul trezorier al Partidului Muncitorilor, aflat la putere รฎn Brazilia, ศi Renato de Souza Duque, fostul preศedinte al serviciilor pentru รฎntreprinderi ale Petrobras." } }
{ "translation": { "en": "Dirceu is the most senior member of the ruling Workers' Party to be taken into custody in connection with the scheme. Dirceu served as former President Luiz Inacio Lula da Silva's chief of staff between 2003 and 2005. He was arrested early August in his home, where he already was under house arrest serving an 11-year sentence for his involvement in a cash-for-votes scheme in Congress more than 10 years ago. Prosecutors have said that Dirceu masterminded the kickback scheme at Petrobras, accepted bribes while in office and continued to receive payments from contractors after he was jailed in late 2013 for the vote-buying scandal.", "ro": "Dirceu este cel mai vechi membru al Partidului Muncitorilor aflat la guvernare luat รฎn custodie pentru legฤturile cu aceastฤ schemฤ. Dirceu a servit ca ศef de cabinet al fostului preศedinte Luiz Inacio Lula da Silva รฎntre 2003 ศi 2005. A fost arestat la รฎnceputul lui august de acasฤ, unde deja se afla sub arest la domiciliu, cu o pedeapsฤ de 11 ani pentru implicarea รฎntr-o schemฤ de cumpฤrare a voturilor รฎn Congres cu peste 10 ani รฎn urmฤ. Procurorii au declarat cฤ Dirceu a dezvoltat schema de luare de mitฤ de la Petrobras, a acceptat mitฤ รฎn timp ce se afla รฎn funcศie ศi a continuat sฤ primeascฤ plฤศi de la antreprenori dupฤ ce a fost รฎnchis la sfรขrศitul lui 2013 pentru scandalul voturilor cumpฤrate." } }
{ "translation": { "en": "According to prosecutors, the scheme at Petrobras involved roughly $2 billion in bribes and other illegal funds. Some of that money was allegedly funneled back to campaign coffers of the ruling party and its allies. It also allegedly included the payment of bribes to Petrobras executives in return for inflated contracts. 'Miraculous' recovery for Peshawar massacre schoolboy A teenager paralysed after being shot four times in Pakistan's deadliest terror attack has made a \"miraculous\" recovery following treatment in the UK. Muhammad Ibrahim Khan, 13, had been told by doctors in Pakistan that he would never walk again.", "ro": "Conform procurorilor, schema de la Petrobras a implicat aproximativ 2 miliarde de dolari sub formฤ de mitฤ ศi alte fonduri ilegale. O parte din acei bani s-ar fi รฎntors รฎn fondul de campanie al partidului aflat la guvernare ศi al aliaศilor acestora. De asemenea, ar fi inclus mitฤ cฤtre directorii Petrobras รฎn schimbul unor contracte umflate. Recuperarea โmiraculoasฤโ a unui elev supravieศuitor al masacrului de la Peshawar Un adolescent paralizat dupฤ ce fusese รฎmpuศcat de patru ori รฎn cel mai cumplit atac terorist din Pakistan a reuศit o recuperare โmiraculoasฤโ dupฤ ce a urmat un tratament รฎn Regatul Unit. Lui Mohamed Ibrahim Khan, รฎn vรขrstฤ de 13 ani, doctorii din Pakistan รฎi spuseserฤ cฤ nu va mai putea sฤ meargฤ niciodatฤ." } }
{ "translation": { "en": "At least 140 people, mostly children, were killed when gunmen stormed Peshawar's Army Public School last December. Muhammad, who arrived in London last month for surgery, is being discharged from hospital later. Exactly nine months ago, on an ordinary Tuesday morning, Muhammad sat in his first aid class listening to his teachers intently. At the same time seven gunmen disguised in security uniforms were entering the Army Public School. They were strapped with explosives and had one simple mission in mind: Kill every man, woman and child they came across. \"I can't forget what happened that day,\" Muhammad says with a severe stare.", "ro": "Cel puศin 140 de persoane, majoritatea copii, au fost ucise cรขnd bฤrbaศi รฎnarmaศi au atacat ศcoala publicฤ a armatei din Peshawar รฎn luna decembrie a anului trecut. Mohamed, care a sosit la Londra luna trecutฤ pentru operaศie, va fi externat mai tรขrziu din spital. Exact cu nouฤ luni รฎn urmฤ, รฎntr-o dimineaศฤ obiศnuitฤ de marศi, Mohamed stฤtea la ora de primul ajutor ศi รฎศi asculta atent profesorii. Chiar atunci, ศapte bฤrbaศi รฎnarmaศi deghizaศi รฎn uniformele agenศilor de pazฤ intrau รฎn ศcoala publicฤ a armatei. Purtau centuri cu explozivi ศi aveau de รฎndeplinit o misiune simplฤ: sฤ รฎi ucidฤ pe toศi bฤrbaศii, femeile ศi copiii care le ieศeau รฎn cale. โNu pot uita ce s-a รฎntรขmplat รฎn acea ziโ, spune Mohamed cu o privire asprฤ." } }
{ "translation": { "en": "We were sitting in the auditorium, we were asking questions... and then we heard heavy gunfire outside. The terrorists moved inside and they started killing - our teacher was burned alive. Muhammad described pulling four other pupils out of the auditorium as the carnage unfolded. He said he then heard his friend, Hamza calling to him. He said, 'oh brother save me'. I held his hand. That's when I was shot in the back, and he was shot in the head. Most of the people killed in the attack were pupils Hamza died in Muhammad's arms. Muhammad recalled blacking out after that, and the next thing he knew he was in a hospital bed, paralysed from the waist down.", "ro": "Stฤteam รฎn amfiteatru, puneam รฎntrebฤri... apoi am auzit focuri de armฤ afarฤ. Teroriศtii au intrat รฎnฤuntru ศi au รฎnceput sฤ ucidฤ. Profesorul nostru a fost ars de viu. Mohamed descrie cum a scos patru elevi din amfiteatru รฎn timp ce se desfฤศura carnagiul. Apoi spune cฤ ศi-a auzit prietenul, pe Hamza, strigรขndu-l. Spunea โoh, frate, salveazฤ-mฤโ. L-am ศinut de mรขnฤ. Atunci eu am fost รฎmpuศcat รฎn spate, iar el รฎn cap. Cei mai mulศi dintre cei uciศi รฎn atac erau elevi Hamza a murit รฎn braศele lui Mohamed. Mohamed รฎศi aminteศte cฤ imediat dupฤ asta a leศinat ศi cฤ urmฤtorul lucru pe care l-a ศtiut a fost cฤ se afla pe un pat de spital, paralizat de la brรขu รฎn jos." } }
{ "translation": { "en": "Doctors in Peshawar in northern Pakistan, and then Rawalpindi, close to the capital, told his family there was no treatment, and he would never walk again. \"Seeing him I felt like my soul had left my body,\" says Muhammad's father, Sher Khan Those nine months were the hardest in my life. But Mr Khan and his wife, Sherbano, refused to believe that their cricket-mad son would never be able to use his legs again. They campaigned, and appealed for help on Pakistani TV, gaining the support of high profile people such as cricketer turned politician Imran Khan.", "ro": "Doctorii din Peshawar din nordul Pakistanului, apoi cei din Rawalpindi, aproape de capitalฤ, i-au spus familiei sale cฤ nu exista tratament ศi cฤ nu va mai putea merge niciodatฤ. โCรขnd l-am vฤzut, am simศit cum รฎmi iese sufletulโ, spune Sher Khan, tatฤl lui Mohamed. Acele nouฤ luni au fost cele mai grele din viaศa mea. รnsฤ Khan ศi soศia lui, Sherbano, au refuzat sฤ creadฤ cฤ fiul lor atรขt de pasionat de crichet nu-ศi va mai putea folosi vreodatฤ picioarele. Au fฤcut o campanie ศi au cerut ajutor de la televiziunea pakistanezฤ, atrฤgรขnd sprijinul unor oameni faimoศi precum Imran Khan, jucฤtor de crichet devenit politician." } }
{ "translation": { "en": "Finally, they were able to raise the funds to bring Muhammad to the UK and provide him with treatment at London's private Harley Street Clinic. Consultant neurosurgeon Irfan Malik described Muhammad as \"terrified\" when he first arrived at the hospital. \"He'd spent the last [few] months lying on a bed, unable to move side to side,\" says Mr Malik. He was weak, he had a pressure sore on his back. He wasn't in great shape. A vertebra at the base of Muhammad's spine was destroyed Muhammad was shot in his shoulder, his hip, and his back during the attack, damaging his lower spine - leading to paralysis.", "ro": "รntr-un final, au reuศit sฤ strรขngฤ fonduri pentru a-l duce pe Mohamed รฎn Regatul Unit ศi a-i oferi tratament la clinica privatฤ Harley Street din Londra. Neurochirurgul consultant Irfan Malik l-a descris pe Mohamed drept โรฎnspฤimรขntatโ cรขnd acesta a ajuns la spital. โรศi petrecuse ultimele [cรขteva] luni zฤcรขnd รฎn pat, fฤrฤ sฤ se poatฤ miศca de pe o parte pe alta, spune Malik. Era slฤbit, se pusese multฤ presiune pe spatele lui. Nu era รฎntr-o formฤ prea bunฤ. O vertebrฤ de la baza coloanei vertebrale a lui Mohamed fusese distrusฤ Mohamed fusese รฎmpuศcat รฎn umฤr, รฎn ศold ศi รฎn spate รฎn timpul atacului, iar coloana vertebralฤ inferioarฤ รฎi fusese distrusฤ, ducรขnd la paralizie." } }
{ "translation": { "en": "But during six hours of surgery, Mr Malik and his team were able to reattach nerve endings and reconstruct the damaged part of the spine. Even Mr Malik was surprised at what happened next. Exactly one week after the surgery Muhammad stood up and started taking steps and walking. We were not expecting to get that sort of excellent result. That was miraculous,\" he says. Less than two weeks after his operation, Muhammad is ready to leave hospital and start the long road to recovery. Muhammad has defied the odds and started to walk again He says he wants to build his strength and continue his education in the UK. But he says he is determined to return to Pakistan, join the army and help fight terrorism.", "ro": "รnsฤ, รฎn timpul unei operaศii care a durat ศase ore, Malik ศi echipa lui au reuศit sฤ lege din nou terminaศiile nervoase ศi sฤ reconstruiascฤ partea distrusฤ a coloanei. Chiar ศi Malik a fost surprins de ceea ce s-a รฎntรขmplat รฎn continuare. Exact la o sฤptฤmรขnฤ dupฤ operaศie, Mohamed s-a ridicat ศi a รฎnceput sฤ facฤ paศi ศi sฤ meargฤ. Nu ne aศteptam la un rezultat atรขt de bun. A fost un miracolโ, spune acesta. รn mai puศin de douฤ sฤptฤmรขni de la operaศie, Mohamed este gata sฤ pฤrฤseascฤ spitalul ศi sฤ รฎnceapฤ procesul lung de recuperare. Mohamed a sfidat soarta ศi a รฎnceput sฤ meargฤ din nou Vrea sฤ devinฤ puternic ศi sฤ รฎศi continue studiile รฎn Regatul Unit. รnsฤ este hotฤrรขt sฤ revinฤ รฎn Pakistan, sฤ se รฎnroleze รฎn armatฤ ศi sฤ lupte รฎmpotriva terorismului." } }
{ "translation": { "en": "\"I feel like I have a second chance at life,\" he says as he shows off pictures he's drawn of guns scribbled out next to school books and pens Muhammad grows physically stronger every day but the psychological trauma he continues to endure is unimaginable. \"My anger is not diminishing\" he says. In my school little kids were killed. What was their crime? His mother, wiping a tear from her eye, caressed his head and said: \"I can see my son walking again.\" He'll be able to get on with his normal life. 'Super Voice' 4G service from Three offers better signal Three is making use of a lower frequency 4G spectrum that can travel more widely", "ro": "โSimt cฤ am รฎncฤ o ศansฤ la viaศฤโ spune el, arฤtรขnd imaginile cu arme desenate de el lรขngฤ manuale ศcolare ศi stilouri Fizic, Mohamed devine tot mai puternic รฎn fiecare zi, รฎnsฤ trauma psihologicฤ prin care trece ศi acum este de neimaginat. โFuria mea nu a scฤzutโ, mฤrturiseศte el. รn ศcoala mea au fost uciศi copii mici. Ce crimฤ au comis ei? Mama lui รฎศi ศterge o lacrimฤ, รฎl mรขngรขie pe creศtet ศi spune: โรmi vฤd fiul mergรขnd din nouโ. Va putea sฤ-ศi continue firesc viaศa. Serviciul 4G โSuper Voiceโ de la Three oferฤ semnal mai bun Three foloseศte un spectru 4G cu o frecvenศฤ mai joasฤ, care poate acoperi o zonฤ mai extinsฤ" } }
{ "translation": { "en": "Mobile phone provider Three has launched a UK service it says will improve reception inside buildings and in rural black spots. Its 4G Super Voice enables customers to make calls and send texts using a lower frequency spectrum. Other networks are looking into introducing the technology, known as Voice Over Long-Term Evolution (VoLTE). It currently works on only the Samsung Galaxy S5, but recent iPhone handsets will be added in the coming months. Three said up to 5.5 million customers would have access to the service by 2017.", "ro": "Furnizorul de telefonie mobilฤ Three a lansat รฎn Regatul Unit un serviciu despre care spune cฤ va รฎmbunฤtฤศi recepศia รฎn interiorul clฤdirilor ศi รฎn zonele rurale fฤrฤ semnal. Serviciul 4G Super Voice le permite clienศilor sฤ efectueze apeluri ศi sฤ trimitฤ mesaje text folosind un spectru cu o frecvenศฤ mai joasฤ. ศi alte reศele intenศioneazฤ sฤ introducฤ aceeaศi tehnologie, cunoscutฤ ca โVoice Over Long-Term Evolution (VoLTE)โ. Aceasta funcศioneazฤ momentan doar cu Samsung Galaxy S5, รฎnsฤ telefoanele iPhone recente vor beneficia de ea รฎn lunile urmฤtoare. Three menศioneazฤ cฤ pรขnฤ la 5,5 milioane de clienศi vor avea acces la serviciu pรขnฤ รฎn 2017." } }
{ "translation": { "en": "Chief technology officer Bryn Jones said: \"By the end of the year, one million of our customers will have access to better indoor coverage and be able to use their phones in more places than ever before.\" Stars prepare for panto season Pantomime season is big business for theatres up and down the UK, with many getting ready for this year's season now. Some of the biggest names in showbusiness now take part in the yuletide theatre. Matthew Kelly and Hayley Mills will be appearing in Cinderella - one as an ugly sister, the other as fairy godmother. They reveal their panto secrets to BBC Breakfast. Steven Wilson: 'If I don't do anything, I feel this creeping guilt'", "ro": "Responsabilul ศef pentru tehnologie, Bryn Jones a declarat: โPรขnฤ la sfรขrศitul anului, un milion dintre clienศii noศtri vor avea acces la o acoperire mai bunฤ รฎn interior ศi รฎศi vor putea folosi telefoanele รฎn mai multe locuri ca pรขnฤ acumโ. Vedetele se pregฤtesc pentru stagiunea de pantomimฤ Stagiunea de pantomimฤ este foarte importantฤ pentru teatrele din tot Regatul Unit, multe dintre ele pregฤtindu-se acum pentru stagiunea din acest an. Acum, la teatrul de Crฤciun participฤ unele dintre numele cele mai mari din showbusiness. Matthew Kelly ศi Hayley Mills vor apฤrea รฎn Cenuศฤreasa - primul รฎn rolul uneia dintre surorile rele, iar a doua รฎn rolul zรขnei. Aceศtia dezvฤluie secretele pantomimei lor la BBC Breakfast. Steven Wilson: โDacฤ nu fac nimic, mฤ simt vinovatโ" } }
{ "translation": { "en": "Steven Wilson was recently the big winner at the Progressive Music Awards Steven Wilson is often dubbed the hardest working musician in the world of progressive rock. The multi-talented musician won three prizes at this month's Progressive Music Awards in London, including album of the year for Hand. The Guardian's five-star review called it \"a smart, soulful and immersive work of art.\" Since the 1980s, Wilson has been the driving force in a number of musical projects, the best known of which is the rock band Porcupine Tree. Now, ahead of two sell-out shows at the Royal Albert Hall, Wilson is releasing a vinyl-only double LP, Transience, to showcase the \"more accessible\" side of his solo output.", "ro": "Steven Wilson a fost desemnat recent drept marele cรขศtigฤtor al Progressive Music Awards Steven Wilson a fost numit de multe ori drept cel mai muncitor muzician din lumea rockului progresiv. Talentatul muzician a cรขศtigat trei premii la Progressive Music Awards, care a avut loc luna aceasta la Londra, printre care ศi premiul pentru cel mai bun album al anului pentru Hand. รn recenzia sa de cinci stele, The Guardian a numit albumul โo operฤ de artฤ inteligentฤ, expresivฤ ศi captivantฤโ. รncฤ din anii 1980, Wilson este motorul mai multor proiecte muzicale, cel mai cunoscut dintre acestea fiind trupa de rock Porcupine Tree. Acum, รฎnainte de douฤ spectacole cu casa รฎnchisฤ la Royal Albert Hall, Wilson lanseazฤ un dublu LP doar รฎn format vinil, Transience, pentru a arฤta latura โmai accesibilฤโ a activitฤศii sale solo." } }
{ "translation": { "en": "He tells the BBC about his love of vinyl, his busy schedule and explains how comic actor Matt Berry came to be his support act. What does vinyl mean to you? I grew up at the very tail end of the vinyl era, and at the time, I remember, we couldn't wait for CD to come along because vinyl was so frustrating. You would buy the record, take it home, and it would have a scratch, and you would have to take it back again. I love CDs, and for some kinds of music - classical for example - it is better than vinyl. But the problem with the CD and digital downloads is that there's nothing you can really cherish or treasure. Owning vinyl is like having a beautiful painting hanging in your living room.", "ro": "A povestit pentru BBC despre dragostea lui pentru viniluri ศi despre programul sฤu รฎncฤrcat ศi a explicat cum a ajuns actorul de comedie Matt Berry sฤ รฎi deschidฤ spectacolele. Ce รฎnseamnฤ vinil pentru tine? Am crescut chiar รฎn perioada de sfรขrศit a erei vinilurilor ศi รฎmi amintesc cฤ atunci abia aศteptam apariศia CD-ului, cฤci vinilul era atรขt de enervant. Cumpฤrai un disc, mergeai cu el acasฤ, avea o zgรขrieturฤ ศi trebuia sฤ รฎl aduci รฎnapoi. Iubesc CD-urile, iar pentru anumite tipuri de muzicฤ, de exemplu cea clasicฤ, sunt mai bune decรขt vinilurile. รnsฤ problema cu CD-urile ศi cu descฤrcฤrile digitale este aceea cฤ nu mai existฤ nimic pe care sฤ รฎl preศuieศti cu adevฤrat. Sฤ ai un vinil e ca ศi cum ai avea un tablou frumos agฤศat รฎn sufragerie." } }
{ "translation": { "en": "It's something you can hold, pore over the lyrics and immerse yourself in the art work. I thought it was just a nostalgic thing, but it can't be if kids too young to remember vinyl are enjoying that kind of experience. Do you have a piece of vinyl that you treasure? The truth is I got rid of 100% of my vinyl in the 90s. All the vinyl I have is re-bought. I started off from the perspective that I wanted to recreate the collection I had when I was 15, but it's gone beyond that. The first record which I persuaded my parents to buy for me was Electric Light Orchestra's Out of the Blue.", "ro": "E ceva ce poศi ศine รฎn mรขnฤ, รฎn timp ce te laศi absorbit de versuri ศi copleศit de actul artistic. Am crezut cฤ e doar o chestie nostalgicฤ, รฎnsฤ nu are cum sฤ fie aศa dacฤ unor puศti prea tineri sฤ-ศi aminteascฤ de viniluri le place acest gen de experienศฤ. Ai vreun vinil la care ศii รฎn mod special? Recunosc cฤ am scฤpat de toate vinilurile รฎn anii '90. Toate vinilurile pe care le am sunt cumpฤrate din nou. Am pornit de la ideea de a reface colecศia pe care o aveam la 15 ani, รฎnsฤ am trecut de limita aceea. Primul disc pe care mi-am convins pฤrinศii sฤ mi-l cumpere a fost Out of the Blue de la Electric Light Orchestra." } }
{ "translation": { "en": "If I still had my original copy, it would have sentimental value, but, alas, it's in a charity shop somewhere. Steven Wilson hopes the album will be a doorway for potential new fans Why release your new compilation Transience on vinyl? It was originally conceived as an idea for Record Store Day, but we missed the boat on that. My record company had suggested I put together some of my shorter, more accessible songs. I got a bit obsessed by the idea to make something like \"an introduction to Steven Wilson,\" and I was committed to it being a vinyl-only release. Anyone who buys the vinyl does also get a high-resolution download.", "ro": "Dacฤ aศ mai fi avut รฎncฤ exemplarul iniศial, acesta ar fi avut valoare sentimentalฤ, รฎnsฤ, din pฤcate, se aflฤ pe undeva printr-un magazin de caritate. Steven Wilson sperฤ cฤ albumul va fi o poartฤ cฤtre posibili fani noi De ce ศi-ai lansat noua compilaศie Transience pe vinil? Aceasta a fost conceputฤ iniศial ca idee pentru Ziua magazinelor de discuri, รฎnsฤ am ratat ocazia. Casa mea de discuri sugerase sฤ adun cรขteva dintre melodiile mele mai scurte ศi mai accesibile. Am ajuns sฤ fiu uศor obsedat de ideea de a face ceva gen โintroducere รฎn muzica lui Steven Wilsonโ ศi am ศinut neapฤrat ca proiectul sฤ fie lansat doar pe vinil. Cine cumpฤrฤ vinilul primeศte, de asemenea, ศi o variantฤ descฤrcatฤ la rezoluศie รฎnaltฤ." } }
{ "translation": { "en": "Do you have a concern that the album won't show your work in a true light?", "ro": "Eศti รฎngrijorat cฤ albumul nu va arฤta muzica ta รฎn adevฤrata ei luminฤ?" } }
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/wmt_en_ro/test.json
|
{ "translation": { "en": "UN Chief Says There Is No Military Solution in Syria Secretary-General Ban Ki-moon says his response to Russia's stepped up military support for Syria is that \"there is no military solution\" to the nearly five-year conflict and more weapons will only worsen the violence and misery for millions of people. The U.N. chief again urged all parties, including the divided U.N. Security Council, to unite and support inclusive negotiations to find a political solution. Ban told a news conference Wednesday that he plans to meet with foreign ministers of the five permanent council nations - the U.S., Russia, China, Britain and France - on the sidelines of the General Assembly's ministerial session later this month to discuss Syria.", "ro": "ศeful ONU declarฤ cฤ nu existฤ soluศii militare รฎn Siria Secretarul General Ban Ki-moon afirmฤ cฤ rฤspunsul sฤu la suportul militar al Rusiei pentru Siria este cฤ โnu existฤ o soluศie militarฤโ la conflictul care dureazฤ de aproape cinci ani iar mai multe arme nu ar face decรขt sฤ agraveze violenศa ศi suferinศa a milioane de oameni. ศeful ONU a solicitat din nou tuturor pฤrศilor, inclusiv Consiliului de securitate ONU divizat sฤ se unifice ศi sฤ susศinฤ negocierile pentru a gฤsi o soluศie politicฤ. Ban a declarat miercuri รฎn cadrul unei conferinศe cฤ intenศioneazฤ sฤ se รฎntรขlneascฤ luna aceasta cu miniศtrii de externe din cinci ศฤri permanent prezente รฎn consiliu - SUA, Rusia, China, Anglia ศi Franศa - pe marginea sesiunii ministeriale a Adunฤrii Generale pentru a discuta despre Siria." } }
{ "translation": { "en": "He expressed regret that divisions in the council and among the Syrian people and regional powers \"made this situation unsolvable.\" Ban urged the five permanent members to show the solidarity and unity they did in achieving an Iran nuclear deal in addressing the Syria crisis. 8 Poll Numbers That Show Donald Trump Is For Real Some have tried to label him a flip-flopper. Others have dismissed him as a joke. And some are holding out for an implosion. But no matter how some Republicans are trying to drag Donald Trump down from atop the polls, it hasn't worked (yet).", "ro": "Ban ศi-a exprimat regretul cฤ divizฤrile รฎn consiliu ศi รฎntre poporul sirian ศi puterile regionale โau fฤcut aceastฤ situaศie de nerezolvatโ. Ban le-a cerut celor cinci membri permanenศi sฤ dea dovadฤ de solidaritatea ศi unitatea arฤtate atunci cรขnd au reuศit sฤ รฎncheie un acord referitor la armele nucleare ale Iranului, abordรขnd astfel criza din Siria. 8 cifre din sondaje care aratฤ cฤ Donald Trump are ศanse reale Unii au รฎncercat sฤ รฎl eticheteze ca politician โflip-flopโ. Alศii l-au numit o glumฤ. Iar alศii aศteaptฤ implozia. รnsฤ indiferent de modul รฎn care unii republicani รฎncearcฤ sฤ รฎl dฤrรขme pe Donald Trump din vรขrful sondajelor, nu a funcศionat (รฎncฤ)." } }
{ "translation": { "en": "Ten of the last 11 national polls have shown Donald Trump's lead at double digits, and some are starting to ask seriously what it means for the real estate mogul's nomination chances. Of course, it's still early in the election cycle. None of this is to say that Trump is likely to win the Republican nomination. Pundits point out that at this time in 2011, Rick Perry's lead was giving way to a rising Herman Cain, neither of whom won even one state in the nomination process. And there are many reasons he would struggle in a general election. But outside groups like Jeb Bush's Super PAC and the economic conservative group Club for Growth are recognizing Trump's staying power and beginning to unload their dollars to topple him.", "ro": "Zece din ultimele 11 sondaje naศionale au arฤtat cฤ Donald Trump conduce cu un procent din douฤ cifre iar unele voci รฎncep sฤ se รฎntrebe serios ce รฎnseamnฤ acest lucru pentru ศansele de numire ale mogulului imobiliar. Desigur, este รฎncฤ prematur. Nimic din toate acestea nu spune cฤ Trump va cรขศtiga cursa pentru nominalizarea republicanilor. Pundits aratฤ cฤ, รฎn aceeaศi perioadฤ a anului 2011, avansul lui Rick Perry รฎi fฤcea loc lui Herman Cain รฎn sondaje, dar niciunul dintre ei nu a cรขศtigat รฎn vreun stat รฎn cursa de nominalizare. Iar motivele pentru care s-ar lupta din greu la alegerile generale sunt numeroase. รnsฤ grupurile din exterior precum Super PAC al lui Jeb Bush ศi grupul conservator economic Club for Growth admit puterea lui Trump ศi รฎncep sฤ รฎl susศinฤ cu bani." } }
{ "translation": { "en": "Here are some recent poll numbers that suggest that the real estate mogul isn't just a passing phase: Trump's favorability ratings have turned 180 degrees. Right before Donald Trump announced his candidacy in mid-June, a Monmouth University poll showed only two in 10 Republicans had a positive view of the real estate mogul. By mid-July, it was 40 percent. In early August, it was 52 percent. Now, six in 10 Republicans have a favorable view of Donald Trump. Roughly three in 10 say they have a negative view. And these numbers hold up in early states. A Quinnipiac poll in Iowa last week found that 60 percent of Republicans there had a favorable view of Trump.", "ro": "รn continuare vฤ prezentฤm cรขteva cifre din sondaje recente care sugereazฤ cฤ mogulul imobiliar nu este doar ceva trecฤtor: Cifrele care indicฤ susศinerea faศฤ de Trump s-au รฎntors la 180 grade. Chiar รฎnainte ca Donald Trump sฤ รฎศi anunศe candidatura, la mijlocul lui iunie, un sondaj realizat de Universitatea din Monmouth arฤta cฤ doar doi din 10 republicani aveau o pฤrere pozitivฤ despre mogulul imobiliar. Pรขnฤ la mijlocul lui iulie, procentul a urcat la 40%. La รฎnceputul lui august, era 52%. รn prezent, ศase din 10 republicani au o pฤrere favorabilฤ despre Donald Trump. Aproximativ trei din 10 declarฤ cฤ au o pฤrere negativฤ. Aceste cifre se menศin. Un sondaj realizat sฤptฤmรขna trecutฤ de Quinnipiac รฎn Iowa a concluzionat cฤ 60% dintre republicanii din regiune au o pฤrere favorabilฤ despre Trump." } }
{ "translation": { "en": "Two-thirds of GOP voters would be happy with Trump as the nominee. In a CNN/ORC poll last week, 67 percent of Republicans said they would be either \"enthusiastic\" or \"satisfied\" if Trump were the nominee. Only two in 10 say they would be \"upset\" if he were the nominee. Only Ben Carson generates roughly the same level of enthusiasm as Trump (43 percent say they would be \"enthusiastic\" vs. 40 percent who say the same of Trump). The next closest in enthusiasm? Marco Rubio with only 21 percent.", "ro": "Douฤ treimi dintre alegฤtorii GOP ar fi fericiศi dacฤ Trump ar cรขศtiga cursa pentru nominalizare. รntr-un sondaj realizat sฤptฤmรขna trecutฤ de CNN/ORC, 67% dintre republicani au declarat cฤ ar fi โentuziasmaศiโ sau โmulศumiศiโ dacฤ Trump ar cรขศtiga cursa pentru nominalizare. Doar doi din 10 declarฤ cฤ ar fi โsupฤraศiโ dacฤ Trump ar cรขศtiga cursa pentru nominalizare. Doar Ben Carson genereazฤ aproximativ acelaศi nivel de entuziasm ca Trump (43% declarฤ cฤ ar fi โentuziasmaศiโ faศฤ de 40% care declarฤ acelaศi lucru despre Trump). Cel mai aproape รฎn ceea ce priveศte entuziasmul? Marco Rubio, cu doar 21%." } }
{ "translation": { "en": "On the flip side, 47 percent of Republican voters say they would be \"dissatisfied\" or \"upset\" if establishment favorite Jeb Bush becomes the nominee. A majority of Republicans don't see Trump's temperament as a problem. While Donald Trump has been widely criticized for his bombast and insults, 52 percent of leaned Republican voters nationwide think that the real estate mogul has the right temperament to be president, according to Monday's ABC News/Washington Post poll. The same number holds in the first-in-the-nation caucus state of Iowa, where the same 52 percent of Republicans think he has the personality to be commander in chief, according to Quinnipiac last week.", "ro": "De partea cealaltฤ, 47% dintre alegฤtorii republicani afirmฤ cฤ ar fi โnemulศumiศiโ sau โsupฤraศiโ dacฤ favoritul Jeb Bush cรขศtigฤ cursa pentru nominalizare. Majoritatea republicanilor nu considerฤ temperamentul lui Trump o problemฤ. Deศi Donald Trump a fost puternic criticat pentru insultele aduse ศi stilul sฤu bombastic, 52% dintre alegฤtorii republicani la nivel naศional considerฤ cฤ mogulul imobiliar are temperamentul potrivit pentru a fi preศedinte, conform sondajului realizat luni de ABC News/Washington Post. Regฤsim aceleaศi cifre รฎn statul Iowa, unde tot 52% dintre republicani cred cฤ Trump are personalitatea potrivitฤ pentru a fi conducฤtor, conform sondajului realizat sฤptฤmรขna trecutฤ de Quinnipiac." } }
{ "translation": { "en": "Still, 44 percent think he doesn't have the personality to serve effectively, and almost six in 10 independents say his temperament does not belong in the White House, according to ABC/Post. Republican voters are getting used to the idea. When they put on their pundit hats, Republican voters think Trump is for real. When asked who is most likely to win the GOP nomination, four in 10 said Trump was the best bet, according to a CNN/ORC poll out last week. That's a change from when four in 10 placed their money on Jeb Bush in late July. Full disclosure: GOP voters haven't had the clearest crystal ball in the past.", "ro": "Totuศi, 44% sunt de pฤrere cฤ nu are personalitatea necesarฤ pentru a acศiona eficient ศi aproape ศase din 10 independenศi afirmฤ cฤ temperamentul sฤu nu are ce cฤuta la Casa Albฤ, conform ABC/Post. Alegฤtorii republicani se obiศnuiesc cu ideea. Atunci cรขnd iau atitudinea de intelectuali, alegฤtorii republicani considerฤ cฤ Trump este autentic. Conform unui sondaj realizat sฤptฤmรขna trecutฤ de CNN/ORC, la รฎntrebarea cine are cele mai multe ศanse sฤ cรขศtige cursa pentru nominalizare GOP, patru din 10 au declarat cฤ Trump. Situaศia s-a schimbat faศฤ de finalul lui iulie, cรขnd patru din 10 ar fi pariat pe Jeb Bush. Informare completฤ: รฎn trecut, alegฤtorii GOP nu au citit foarte bine viitorul." } }
{ "translation": { "en": "At this time last cycle, four in 10 Republicans picked Rick Perry to win the nomination, vs. only 28 percent for eventual nominee Mitt Romney. Still, it shows that a plurality of GOP voters see Trump's campaign as plausible. Even if Republicans rallied around another candidate, Trump still beats almost everyone. Some pundits point out that the splintered field is likely contributing to Trump's lead, while anti-Trump support is be spread diffusely among more than a dozen other candidates. But a Monmouth University poll in early September shows that, in a hypothetical head-to-head matchup between Trump and most other Republican candidates, Trump almost always garners majority support.", "ro": "รn aceeaศi perioadฤ a ultimelor alegeri, patru din 10 republicani l-au ales pe Rick Perry รฎn cursa pentru nominalizare, faศฤ de doar 28% pentru Mitt Romney. รnsฤ, aceste cifre aratฤ cฤ majoritatea alegฤtorilor GOP considerฤ plauzibilฤ campania lui Trump. Chiar dacฤ republicanii sau repliat spre un alt candidat. Trump รฎncฤ se aflฤ รฎn fruntea tuturor. Unele voci spun cฤ situaศia divizatฤ va contribui probabil la victoria lui Trump, รฎn timp ce susศinerea contra lui Trump se va รฎmpฤrศi la mai mult de doisprezece candidaศi. รnsฤ un sondaj derulat la รฎnceputul lui septembrie de Universitatea din Monmouth aratฤ cฤ, รฎn situaศia ipoteticฤ a unei colaborฤri รฎntre Trump ศi majoritatea celorlalศi candidaศi republicani, aproape รฎntotdeauna Trump va beneficia de susศinerea majoritarฤ." } }
{ "translation": { "en": "He leads Carly Fiorina by 13 points, Marco Rubio by 14 points, Walker by 15 points, Jeb Bush by 19 points, and, finally, Rand Paul, John Kasich and Chris Christie by 33 points each. He's in a dead heat with Ted Cruz. The only candidate who beats him? Ben Carson would lead the businessman by a wide 19 points in a hypothetical head-to-head. A bare majority of Donald Trump's supporters say they've made up their minds. A new CBS/NYT poll out on Tuesday shows that just more than half of voters who support Trump say they have locked in their votes. Obviously, a lot can happen to change that, and no one can really say they would never change their mind.", "ro": "Trump se aflฤ la distanศฤ de 13 puncte de Carly Fiorina, la 14 puncte de Marco Rubio, la 15 puncte de Walker, la 19 puncte de Jeb Bush ศi, รฎn cele din urmฤ, la cรขte 33 de puncte faศฤ de Rand Paul, John Kasich ศi Chris Christie. Este aproape la egalitate cu Ted Cruz. Singurul candidat care รฎl รฎnvinge? Ben Carson l-ar รฎnvinge pe omul de afaceri cu 19 puncte รฎntr-o confruntare ipoteticฤ de unu la unu. Majoritatea susศinฤtorilor lui Donald Trump declarฤ cฤ s-au decis. Un nou sondaj realizat marศi de CBS/NYT aratฤ cฤ peste jumฤtate dintre alegฤtorii care รฎl susศin pe Trump declarฤ cฤ nu รฎศi schimbฤ opศiunea de vot. Evident, se pot รฎntรขmpla multe รฎn acest sens ศi nimeni nu poate spune cฤ aceศtia nu se vor rฤzgรขndi niciodatฤ." } }
{ "translation": { "en": "46 percent said they are leaving the door open to switching candidates. Still, Trump's strongest competition at the moment is from fellow outsider neurosurgeon Ben Carson, but voters who say they have made up their minds are twice as likely to go for Trump. Six in 10 Republicans say they agree with Trump on immigration. Even since Donald Trump called immigrants from Mexico \"rapists\" in his campaign announcement speech two months ago, immigration has been front and center in the 2016 conversation. Some are worried that Trump's bombast will drive crucial Hispanic voters away from the Republican Party and damage rebranding efforts.", "ro": "46% afirmฤ cฤ lasฤ portiศa deschisฤ posibilitฤศii de a-ศi schimba opศiunea. Cu toate acestea, cel mai important adversar al lui Trump este รฎn prezent neurochirurgul Ben Carson, รฎnsฤ este de douฤ ori mai probabil ca alegฤtorii care declarฤ cฤ s-au decis sฤ voteze cu Trump. ศase din 10 republicani afirmฤ cฤ sunt de acord cu Trump รฎn problema imigrฤrii. De cรขnd Donald Trump i-a numit pe imigranศii din Mexic โviolatoriโ รฎn discursul de deschidere a campaniei sale, รฎn urmฤ cu douฤ luni, imigrarea a fost subiectul central รฎn campania pentru 2016. Unii sunt รฎngrijoraศi cฤ stilul bombastic al lui Trump va duce la o scindare รฎntre alegฤtorii hispanici importanศi ศi Partidul Republican ศi va prejudicia eforturile de rebranding." } }
{ "translation": { "en": "But according to Monday's new ABC/Post poll, six in 10 Republicans say they agree with Trump on immigration issues. So as long as immigration remains in the spotlight, it seems Donald Trump will remain too. Frustration with government is climbing to new highs. Donald Trump and Ben Carson now account for roughly half of the support from Republican voters, largely due to their outsider status. Six in 10 Republicans in Monday's new ABC/Post poll say they want a political outsider over someone with government experience. And they are angry at Washington, too.", "ro": "รnsฤ, conform sondajului realizat luni de ABC/Post, ศase din 10 republicani afirmฤ cฤ sunt de acord cu Trump รฎn problema imigrฤrii. Aศa cฤ, se pare cฤ atรขta timp cรขt problema imigrฤrii rฤmรขne รฎn lumina reflectoarelor, la fel va rฤmรขne ศi Doland Trump. Frustrarea faศฤ de autoritฤศi atinge noi culmi. Donald Trump ศi Ben Carson sunt acum susศinuศi de aproape jumฤtate dintre alegฤtorii republicani, รฎn mare parte datoritฤ statutului lor de outsideri. Conform sondajului realizat luni de ABC/Post, ศase din 10 republicani afirmฤ cฤ preferฤ un outsider politic รฎn detrimentul cuiva cu experienศฤ รฎn guvernare. Oamenii sunt de asemenea supฤraศi pe autoritฤศile de la Washington." } }
{ "translation": { "en": "A Des Moines Register/Bloomberg poll in Iowa from two weeks ago shows that three in four Iowa Republicans are frustrated with Republicans in Congress, with 54 percent \"unsatisfied\" and 21 percent \"mad as hell.\" Jeremy Corbyn to make debut at Prime Minister's Questions Since his election, Mr Corbyn's debut at PMQs has been keenly awaited New Labour leader Jeremy Corbyn is to make his debut at Prime Minister's Questions later, taking on David Cameron for the first time.", "ro": "Un sondaj derulat รฎn urmฤ cu douฤ sฤptฤmรขni รฎn Iowa de cฤtre Des Moines Register/Bloomberg aratฤ cฤ trei din patru republicani din Iowa sunt frustraศi de prestaศia republicanilor din COngres, 54% declarรขndu-se โnemulศumiศiโ iar 21% โnervoศi la culmeโ. Jeremy Corbyn รฎศi face debutul la Prime Minister's Questions รncฤ de la alegerea sa, debutul domnului Corbyn la PMQs a fost รฎndelung aศteptat Noul lider al Partidului Laburist, Jeremy Corbyn, รฎศi va face mai tรขrziu debutul la Prime Minister's Questions, confruntรขndu-se pentru prima datฤ cu David Cameron." } }
{ "translation": { "en": "Mr Corbyn will rise to ask the first of his six allotted questions shortly after midday, with his performance likely to be closely scrutinised by the media and Labour MPs. He has called for \"less theatre and more facts\" at the weekly showpiece. He has also said he could skip some sessions, leaving them to colleagues. The encounter will be the first parliamentary test of Mr Corbyn's leadership, coming after his appointment of a shadow cabinet and his speech to the TUC annual congress on Tuesday.", "ro": "Dl Corbyn va adresa primele dintre cele ศase รฎntrebฤri la care are dreptul la scurt timp dupฤ prรขnz; prestaศia sa va fi probabil analizatฤ รฎndeaproape de mass-media ศi parlamentarii laburiศti. รn cadrul apariศiilor sฤptฤmรขnale, el a cerut โmai puศin teatru ศi mai multe fapteโ. A declarat de asemenea cฤ poate renunศa la cรขteva participฤri ศi cฤ le cedeazฤ colegilor sฤi. Confruntarea va fi primul test parlamentar al Dl Corbyn รฎn poziศie de lider, venind dupฤ ce a numit un โcabinet fantomฤโ ศi dupฤ discursul pe care l-a ศinut marศi la congresul anual TUC." } }
{ "translation": { "en": "Meanwhile, the Labour leader's decision to stand in silence during the singing of the national anthem at a service on Tuesday to mark the 75th anniversary of the Battle of Britain has attracted criticism from a number of Tory MPs and is the focus of several front page stories in the newspapers. Mr Corbyn's decision not to sing the national anthem has attracted attention A spokesman for Mr Corbyn said he had \"stood in respectful silence\" and did recognise the \"heroism of the Royal Air Force in the Battle of Britain.\"", "ro": "รntre timp, decizia liderului Partidului laburist de a pฤstra tฤcerea la rostirea imnului naศional รฎn cadrul unei slujbe ศinute marศi cu ocazia aniversฤrii a 75 de ani de la Bฤtฤlia Angliei a atras critici din partea unor parlamentari conservatori ศi a ศinut prima paginฤ a ziarelor. Decizia domnului Corbyn de a nu cรขnta imnul naศional a atras atenศia Un purtฤtor de cuvรขnt al Dl Corbyn a declarat cฤ acesta โa pฤstrat tฤcerea รฎn mod respectuosโ ศi a recunoscut โeroismul Forศelor aeriene britanice รฎn Bฤtฤlia Angliei.โ" } }
{ "translation": { "en": "But a member of Mr Corbyn's shadow cabinet, Owen Smith, told BBC Two's Newsnight programme he would have advised the Labour leader to sing the national anthem \"irrespective\" of his belief that the monarchy should be abolished. Nearly a dozen shadow ministers have refused to serve in Mr Corbyn's top team, citing differences over the economy, defence and foreign affairs, while less than a sixth of the parliamentary party originally backed him as leader. BBC political correspondent Robin Brant says policy differences are also \"stacking up\" within Labour following Mr Corbyn's appointment over its position on the European Union and the government's cap on benefits.", "ro": "รnsฤ un membru al cabinetului fantomฤ al Dl Corbyn, Owen Smith, a declarat pentru emisiunea Two's Newsnight transmisฤ de BBC cฤ i-ar fi recomandat liderului laburist sฤ cรขnte imnul naศional โindiferentโ de credinศa sa cฤ monarhia ar trebui abolitฤ. รn jur de doisprezece miniศtri din cabinetul fantomฤ au refuzat sฤ facฤ parte din echipa de frunte a Dl Corbyn, argumentรขnd prin diferenศe de opinie legate de economie, apฤrare ศi externe, รฎn timp ce mai puศin de o ศesime din partidul parlamentar l-a susศinut ca lider. Corespondentul politic al BBC, Robin Brant, declarฤ cฤ diferenศele de politicฤ โse cumuleazฤโ รฎn Partidul Laburist dupฤ numirea domnului Corbyn referitor la poziศia sa faศฤ de Uniunea Europeanฤ ศi limita de beneficii." } }
{ "translation": { "en": "Mr Corbyn told the TUC conference Labour was putting forward amendments to remove the whole idea of a cap altogether. Hours later Mr Smith, the shadow work and pensions secretary, said the party was \"very clear\" that it was only opposing government plans to reduce the level of cap from ยฃ26,000 to ยฃ23,000. Mr Corbyn will be the fifth Labour leader that David Cameron has faced across the despatch box over the past decade since he became Tory leader. The Labour leader, who has promised a different approach to politics, says he has \"crowd sourced\" ideas for questions to ask Mr Cameron and has been given more than 30,000 suggestions.", "ro": "Dl Corbyn a declarat la conferinศa TUC cฤ Partidul Laburist va aduce modificฤri prin care se va elimina integral ideea limitฤrii. Cรขteva ore mai tรขrziu, Dl Smith, Ministrul Muncii ศi Pensiilor, a declarat cฤ partidul โeste foarte clarโ รฎn opoziศia exclusivฤ faศฤ de planurile guvernului de a reduce nivelul โcapโ de la 26.000 lire la 23.000 lire. Dl Corbyn va fi al cincilea lider laburist cu care se confruntฤ David Cameron la tribunฤ รฎn ultimul deceniu, de cรขnd a preluat conducerea Partidului Conservator. Liderul laburist, care a promis o abordare diferitฤ a politicii, spune cฤ are idei โdin surse externeโ pentru รฎntrebฤri pe care sฤ i le adreseze Domnului Cameron ศi cฤ a primit peste 30.000 de sugestii." } }
{ "translation": { "en": "The Islington North MP has said PMQs is too confrontational and that he will refrain from both \"repartee\" and trading barbs, instead vowing to focus on serious issues such as poverty, inequality and the challenges facing young people. Mr Corbyn has said that Angela Eagle, the shadow business secretary, will deputise for him at PMQs when he does not attend - for instance when Mr Cameron is travelling abroad. He has also floated the idea of allowing other colleagues to take the floor on occasion, saying he had approached the Commons Speaker John Bercow to discuss the issue.", "ro": "Parlamentarul Islington North a afirmat cฤ PMQs implicฤ un nivel de confruntare prea รฎnalt ศi cฤ se va abศine de la replici ศi atacuri, angajรขndu-se sฤ se concentreze รฎn schimb pe probleme serioase precum sฤrฤcia, inegalitatea ศi provocฤrile cu care se confruntฤ tinerii. Dl Corbyn a declarat cฤ Angela Eagle, Ministrul de finanศe, รฎi va ศine locul la PMQs atunci cรขnd el nu poate participa - de exemplu atunci cรขnd Dl Cameron se deplaseazฤ รฎn strฤinฤtate. A exprimat de asemenea ideea cฤ va permite altor colegi sฤ ia cuvรขntul ocazional, spunรขnd cฤ l-a abordat pe Preศedintele Camerei Deputaศilor, John Bercow, pentru a discuta acest aspect." } }
{ "translation": { "en": "When he became leader in 2005, Mr Cameron said he wanted to move away from the \"Punch and Judy\" style of politics often associated with PMQs but admitted some years later that he had failed. Since it was first televised in 1990, PMQs has been seen as a key barometer of a leader's judgement, their command of the Commons and their standing among their fellow MPs although critics have argued it has become a caricature and is in need of far-reaching reforms. 'Shot in Joburg': Homeless youth trained as photographers Downtown Johannesburg is a tough place to be homeless.", "ro": "รn 2005, cรขnd a preluat conducerea, Dl Cameron a declarat cฤ doreศte sฤ renunศe la stilul politic โPunch and Judyโ asociat adesea cu PMQs รฎnsฤ a recunoscut cรขศiva ani mai tรขrziu cฤ nu a reuศit รฎn demersul sฤu. De la prima transmisie, รฎn 1990, PMQs a fost consideratฤ un barometru cheie al raศionamentului unui lider, al modului รฎn care acesta conduce Camera Deputaศilor ศi a poziศiei sale รฎn rรขndul colegilor parlamentari, deศi criticii afirmฤ a ca devenit o caricaturฤ ศi cฤ are nevoie de o reformare profundฤ. โCadru รฎn Joburgโ: Tineri fฤrฤ adฤpost beneficiazฤ de cursuri de fotografie Este dificil sฤ fii un om fฤrฤ adฤpost รฎn Johannesburg." } }
{ "translation": { "en": "But one group of former street children have found a way to learn a skill and make a living. \"I was shot in Joburg\" is a non-profit studio that teaches homeless youngsters how to take photographs of their neighbourhood and make a profit from it. BBC News went to meet one of the project's first graduates. JD Sports boss says higher wages could hurt expansion JD Sports Executive Chairman Peter Cowgill says a higher minimum wage for UK workers could mean \"more spending power in the pockets of potential consumers.\" But that spending power is unlikely to outweigh the higher labour costs at his firm, he says.", "ro": "รnsฤ un grup de oameni care au trฤit pe strฤzi รฎn copilฤrie au gฤsit un mod de a รฎnvฤศa o meserie ศi de a-ศi cรขศtiga traiul. โI was shot รฎn Joburgโ este un studio non-profit care รฎi รฎnvaศฤ pe tinerii fฤrฤ adฤpost sฤ facฤ fotografii ale zonelor รฎn care trฤiesc ศi sฤ cรขศtige bani din asta. BBC News s-a รฎntรขlnit cu unul dintre primii absolvenศi ai proiectului. ศeful JD Sports spune cฤ salariile mai mari ar putea dฤuna extinderii Preศedintele JD Sports, Peter Cowgill, declarฤ cฤ o creศtere a salariului minim รฎn Marea Britanie ar putea รฎnsemna โo putere de cumpฤrare mai mare รฎn buzunarele potenศialilor consumatori.โ Este รฎnsฤ puศin probabil ca respectiva putere de cumpฤrare sฤ depฤศeascฤ costurile mai mari pentru forศa de muncฤ รฎn cadrul firmei, afirmฤ el." } }
{ "translation": { "en": "The costs could hit JD Sports' expansion plans, he added, which could mean fewer extra jobs. Thanasi Kokkinakis backed by Tennis Australia president Steve Healy Thanasi Kokkinakis deserves kudos rather than criticism for his behaviour. Thanasi Kokkinakis has been the collateral damage in the recent storm around his friend Nick Kyrgios and deserves kudos rather than criticism for his own behaviour, according to Tennis Australia president Steve Healy.", "ro": "Costurile ar putea avea impact asupra planurilor de extindere ale JD Sports, a adฤugat el, ceea ce ar putea รฎnsemna mai puศine locuri de muncฤ noi. Thanasi Kokkinakis susศinut de preศedintele Tennis Australia, Steve Healy Thanasi Kokkinakis ar merita sฤ fie lฤudat ศi nu criticat pentru comportamentul sฤu. Thanasi Kokkinakis a fost victimฤ colateralฤ รฎn โfurtunaโ creatฤ รฎn jurul prietenului sฤu, Nick Kyrgios, iar comportamentul sฤu meritฤ mai degrabฤ cuvinte de laudฤ ศi nu criticฤ, รฎn opinia preศedintelui Tennis Australia, Steve Healy." } }
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/wmt_en_ro/train.json
|
{ "translation": { "en": "Corrections to votes and voting intentions: see Minutes Assignment conferred on a Member: see Minutes Membership of committees and delegations: see Minutes Decisions concerning certain documents: see Minutes Forwarding of texts adopted during the sitting: see Minutes Dates for next sittings: see Minutes", "ro": "Corectฤrile voturilor ลi intenลฃiile de vot: a se vedea procesul-verbal Misiune รฎncredinลฃatฤ unui deputat: consultaลฃi procesul-verbal Componenลฃa comisiilor ลi a delegaลฃiilor: a se vedea procesul-verbal Decizii privind anumite documente: a se vedea procesul-verbal Transmiterea textelor adoptate รฎn cursul prezentei ลedinลฃe: a se vedea procesul-verbal Calendarul urmฤtoarelor ลedinลฃe: a se vedea procesul-verbal" } }
{ "translation": { "en": "Membership of Parliament: see Minutes Approval of Minutes of previous sitting: see Minutes Membership of Parliament: see Minutes Verification of credentials: see Minutes Documents received: see Minutes Written statements and oral questions (tabling): see Minutes Petitions: see Minutes Texts of agreements forwarded by the Council: see Minutes Action taken on Parliament's resolutions: see Minutes Agenda for next sitting: see Minutes Closure of sitting (The sitting was closed at 7.45 p.m.)", "ro": "Componenลฃa Parlamentului: a se vedea procesul-verbal Aprobarea procesului-verbal al ลedinลฃei precedente: a se vedea procesul-verbal Componenลฃa Parlamentului: a se vedea procesul-verbal Verificarea prerogativelor: a se vedea procesul-verbal Depunere de documente: a se vedea procesul-verbal Declaraลฃii scrise ลi รฎntrebฤri orale (depunere): consultaลฃi procesul-verbal Petiลฃii: a se vedea procesul-verbal Transmiterea de cฤtre Consiliu a textelor acordurilor: a se vedea procesul-verbal Cursul dat rezoluลฃiilor Parlamentului: a se vedea procesul-verbal Ordinea de zi a urmฤtoarei ลedinลฃe: a se vedea procesul-verbal Ridicarea ลedinลฃei (Se levanta la sesiรณn a las 19.45 horas)" } }
{ "translation": { "en": "Election of Vice-Presidents of the European Parliament (deadline for submitting nominations): see Minutes (The sitting was suspended at 12.40 p.m. and resumed at 3.00 p.m.) Election of Quaestors of the European Parliament (deadline for submitting nominations): see Minutes (The sitting was suspended at 3.25 p.m. and resumed at 6.00 p.m.) Agenda for next sitting: see Minutes Closure of sitting (The sitting was closed at 6.15 p.m.) Opening of the sitting (The sitting was opened at 9.35 a.m.) Documents received: see Minutes Approval of Minutes of previous sitting: see Minutes Membership of Parliament: see Minutes", "ro": "Alegerea vicepreลedinลฃilor Parlamentului European (termenul de depunere a candidaturilor): consultaลฃi procesul-verbal (Die Sitzung wird um 12.40 Uhr unterbrochen und um 15.00 Uhr wiederaufgenommen). Alegerea chestorilor Parlamentului European (termenul de depunere a candidaturilor): consultaลฃi procesul-verbal (Die Sitzung wird um 15.25 Uhr unterbrochen und um 18.00 Uhr wiederaufgenommen). Ordinea de zi a urmฤtoarei ลedinลฃe: a se vedea procesul-verbal Ridicarea ลedinลฃei (Die Sitzung wird um 18.15 Uhr geschlossen.) Deschiderea ลedinลฃei (Die Sitzung wird um 9.35 Uhr erรถffnet.) Depunerea documentelor: a se vedea procesul-verbal Aprobarea procesului-verbal al ลedinลฃei precedente: a se vedea procesul-verbal Componenลฃa Parlamentului: a se vedea procesul-verbal" } }
{ "translation": { "en": "Membership of committees (deadline for tabling amendments): see Minutes (The sitting was suspended at 7 p.m. and resumed at 9 p.m.) Agenda for next sitting: see Minutes Closure of sitting (The sitting was suspended at 23.25 p.m.) Documents received: see Minutes Communication of Council common positions: see Minutes (The sitting was suspended at 11.35 a.m. and resumed for voting time at noon) Approval of Minutes of previous sitting: see Minutes Committee of Inquiry into the crisis of the Equitable Life Assurance Society (extension of mandate): see Minutes", "ro": "Componenลฃa comisiilor (termenul de depunere a amendamentelor): consultaลฃi procesul-verbal (La seduta, sospesa alle 19.00, รจ ripresa alle 21.00) Ordinea de zi a urmฤtoarei ลedinลฃe: a se vedea procesul-verbal Ridicarea ลedinลฃei (Die Sitzung wird um 23.25 Uhr geschlossen.) Depunerea documentelor: a se vedea procesul-verbal Comunicarea poziลฃiilor comune ale Parlamentului: a se vedea procesul-verbal (La sรฉance, suspendue ร 11h35 dans l'attente de l'Heure des votes, est reprise ร midi) Aprobarea procesului-verbal al ลedinลฃei precedente: a se vedea procesul-verbal Comisia de anchetฤ privind criza societฤลฃii de asigurฤri \"Equitable Lifeโ (prelungirea mandatului): consultaลฃi procesul-verbal" } }
{ "translation": { "en": "Announcement by the President: see Minutes 1. Membership of committees (vote) 2. Amendment of the ACP-EC Partnership Agreement (vote) 4. Certification of train drivers operating locomotives and trains on the railway system in the Community (vote) 6. Law applicable to non-contractual obligations (\"ROME II\") (vote) 8. Seventh and eighth annual reports on arms exports (vote) Corrections to votes and voting intentions: see Minutes Membership of committees and delegations: see Minutes Request for waiver of parliamentary immunity: see Minutes Decisions concerning certain documents: see Minutes", "ro": "Comunicarea Preลedintelui: consultaลฃi procesul-verbal 1. Componenลฃa comisiilor (vot) 2. Modificarea Acordului de parteneriat ACP-CE (\"Acordul de la Cotonouโ) (vot) 4. Certificarea mecanicilor de locomotivฤ care conduc locomotive ลi trenuri รฎn sistemul feroviar comunitar (vot) 6. Legea aplicabilฤ obligaลฃiilor necontractuale (\"Roma IIโ) (vot) 8. Al ลaptelea ลi al optulea raport anual privind exportul de armament (vot) Corectฤrile voturilor ลi intenลฃiile de vot: a se vedea procesul-verbal Componenลฃa comisiilor ลi a delegaลฃiilor: a se vedea procesul-verbal Cerere de ridicare a imunitฤลฃii parlamentare: consultaลฃi procesul-verbal Decizii privind anumite documente: a se vedea procesul-verbal" } }
{ "translation": { "en": "Written statements for entry", "ro": "Declaraลฃii scrise รฎnscrise" } }
{ "translation": { "en": "Written statements for entry in the register (Rule 116): see Minutes Forwarding of texts adopted during the sitting: see Minutes Dates for next sittings: see Minutes Adjournment of the session I declare the session of the European Parliament adjourned. (The sitting was closed at 1 p.m.) Approval of Minutes of previous sitting: see Minutes Membership of Parliament: see Minutes Request for the defence of parliamentary immunity: see Minutes Appointments to committees (proposal by the Conference of Presidents): see Minutes Documents received: see Minutes Texts of agreements forwarded by the Council: see Minutes", "ro": "Declaraลฃii scrise รฎnscrise รฎn registru (articolul 116 din Regulamentul de procedurฤ): a se vedea procesul-verbal Transmiterea textelor adoptate รฎn cursul prezentei ลedinลฃe: a se vedea procesul-verbal Calendarul urmฤtoarelor ลedinลฃe: a se vedea procesul-verbal รntreruperea sesiunii Dichiaro interrotta la sessione del Parlamento europeo. (La seduta รจ tolta alle 13.00) Aprobarea procesului-verbal al ลedinลฃei precedente: a se vedea procesul-verbal Componenลฃa Parlamentului: a se vedea procesul-verbal Cerere de apฤrare a imunitฤลฃii parlamentare: consultaลฃi procesul-verbal Numiri รฎn comisii (propunerea Conferinลฃei preลedinลฃilor): consultaลฃi procesul-verbal Depunerea documentelor: a se vedea procesul-verbal Transmiterea de cฤtre Consiliu a textelor acordurilor: a se vedea procesul-verbal" } }
{ "translation": { "en": "Action taken on Parliament's resolutions: see Minutes Oral questions and written statements (tabling): see Minutes Written statements (Rule 116): see Minutes Agenda: see Minutes 1. Appointments to parliamentary committees (vote): see Minutes Voting time Agenda for next sitting: see Minutes Closure of sitting (The sitting was closed at 12 midnight) Opening of the sitting (The sitting was opened at 09.05) Documents received: see Minutes Approval of Minutes of previous sitting: see Minutes 1. Protection of passengers against displaced luggage (vote) 2.", "ro": "Continuฤri ale rezoluลฃiilor Parlamentului: consultaลฃi procesul-verbal Declaraลฃii scrise ลi รฎntrebฤri orale (depunere): consultaลฃi procesul-verbal Declaraลฃii scrise (articolul 116 din Regulamentul de procedurฤ) Ordinea de zi: a se vedea procesul-verbal 1. Numiri รฎn comisiile parlamentare (vot): consultaลฃi procesul-verbal Timpul afectat votului Ordinea de zi a urmฤtoarei ลedinลฃe: a se vedea procesul-verbal Ridicarea ลedinลฃei (La seduta รจ tolta alle 24.00) Deschiderea ลedinลฃei (The sitting was opened at 09.05) Depunerea documentelor: a se vedea procesul-verbal Aprobarea procesului-verbal al ลedinลฃei precedente: a se vedea procesul-verbal 1. Protecลฃia pasagerilor รฎmpotriva deplasฤrii bagajelor (vot) 2." } }
{ "translation": { "en": "Approval of motor vehicles with regard to the forward field of vision of the driver (vote) 3. EC-Korea Agreement on scientific and technological cooperation (vote) 4. Mainstreaming sustainability in development cooperation policies (vote) 5. Draft Amending Budget No 1/2007 (vote) 7. EC-Gabon Fisheries Partnership (vote) 10. Limitation periods in cross-border disputes involving personal injuries and fatal accidents (vote) 12. Strategy for a strengthened partnership with the Pacific Islands (vote) 13. The European private company statute (vote) That concludes the vote.", "ro": "Omologarea vehiculelor cu motor cu privire la cรขmpul de vizibilitate รฎnainte al conducฤtorului auto (vot) 3. Acordul CE-Coreea de cooperare ลtiinลฃificฤ ลi tehnologicฤ (vot) 4. Integrarea durabilitฤลฃii รฎn politicile de cooperare pentru dezvoltare (vot) 5. Proiect de buget rectificativ nr.1/2007 (vot) 7. Acordul de parteneriat รฎn domeniul pescuitului รฎntre Comunitatea Europeanฤ ลi Republica Gabonezฤ (vot) 10. Termenele de prescripลฃie aplicabile รฎn cadrul litigiilor transfrontaliere cu privire la vฤtฤmฤrile corporale ลi accidentele mortale (vot) 12. Relaลฃiile UE cu insulele din Pacific: Strategie pentru un parteneriat consolidat (vot) 13. Statutul societฤลฃii private europene (vot) Damit ist die Abstimmungsstunde beendet." } }
{ "translation": { "en": "Corrections to votes and voting intentions: see Minutes Assignment conferred on a Member: see Minutes Membership of committees and delegations: see Minutes Decisions concerning certain documents: see Minutes Forwarding of texts adopted during the sitting: see Minutes Dates for next sittings: see Minutes", "ro": "Corectฤrile voturilor ลi intenลฃiile de vot: a se vedea procesul-verbal Misiune รฎncredinลฃatฤ unui deputat: consultaลฃi procesul-verbal Componenลฃa comisiilor ลi a delegaลฃiilor: a se vedea procesul-verbal Decizii privind anumite documente: a se vedea procesul-verbal Transmiterea textelor adoptate รฎn cursul prezentei ลedinลฃe: a se vedea procesul-verbal Calendarul urmฤtoarelor ลedinลฃe: a se vedea procesul-verbal" } }
{ "translation": { "en": "Written statements for entry", "ro": "Declaraลฃii scrise รฎnscrise" } }
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/xsum/sample.json
|
{"document": "The warning begins at 22:00 GMT on Saturday and ends at 10:00 on Sunday.\nThe ice could lead to difficult driving conditions on untreated roads and slippery conditions on pavements, the weather service warned.\nOnly the southernmost counties and parts of the most westerly counties are expected to escape.\nCounties expected to be affected are Carmarthenshire, Powys, Ceredigion, Pembrokeshire, Denbighshire, Gwynedd, Wrexham, Conwy, Flintshire, Anglesey, Monmouthshire, Blaenau Gwent, Caerphilly, Merthyr Tydfil, Neath Port Talbot, Rhondda Cynon Taff and Torfaen.", "summary": "The Met Office has issued a yellow weather warning for ice across most of Wales."}
{"document": "You can see highlights of Sunderland v Arsenal on Match of the Day at 22:20 BST on Saturday on BBC One and the BBC Sport website.\nStoke and West Ham, for example, have started to climb away from the relegation zone but the biggest worry for Sunderland fans is that their side do not look remotely capable of doing the same.\nI know the Black Cats have got out of trouble before having found themselves in a similar situation but this time, after picking up only two points from their first nine games, things look really desperate for the only top-flight team without a win.\nAt least one element of their struggles seems to be self-inflicted, with everyone at the club feeling sorry for themselves - and not just because they have lost some players to injury and conceded some costly late goals.\nThere is a negative feeling about the place with the manager David Moyes and his players talking about how they have gone backwards since last season, when they should be searching for any kind of spark that could change things around.\nFrom the outside, looking at the way they play and their lack of creativity, it is hard to see what that spark might be or what could fundamentally change under Moyes until the January transfer window opens.\nIf they can get one win under their belt then they will get a bit of belief back but, the longer this winless run goes on, the more negativity there will be.\nMedia playback is not supported on this device\nSunderland finished last season on a high under Sam Allardyce, with a run of just one defeat in their last 11 games securing their safety.\nIn the space of five months, all of that confidence and momentum seems to have been sucked out of the club, despite them effectively having the same group of players who, not so long ago, looked inspired.\nThat is not all down to Moyes, but he has to take some responsibility for it.\nI am yet to see a defined style of play from Sunderland since he took charge at the end of July.\nThat is in contrast to Allardyce's time as manager, when they were resolute and difficult to beat and, at the end of his stint at the Stadium of Light, also played with a purpose when they went forward.\nOff the pitch, Moyes has not helped himself much either.\nThere was no need for him to be so pessimistic when he came out after the second game of the season and announced they would be in a relegation fight, which did not send out the right message to his players or the fans.\nWhen he took charge, he had actually started out by being unrealistically positive - talking about Sunderland becoming a club that regularly finished in the top half of the Premier League - but his expectations went downhill very quickly.\nI know you can argue that he has been proved right, because Sunderland are now battling the drop, but it meant there was a cloud over from them almost as soon as the season had started.\nIt seems to be a case that if you stop Jermain Defoe, you stop Sunderland. His statistics stand up well in comparison to last season, but the rest of their team are not doing enough in attack.\nThey were reliant on Defoe last season too, but others did chip in - in their first nine league games of 2015-16, five players found the net. This time around, only Defoe and Patrick van Aanholt have scored in the same period.\nIt is going to be a massive struggle for them to stay up from the position they are now in anyway, but they badly need a win and quickly. I don't see it coming at home to Arsenal on Saturday, though.\nDo they even look capable of holding out for a draw against the Gunners, the way another struggling team Middlesbrough did at Emirates Stadium last weekend? No.\nIf you struggle to make chances and score goals, as Sunderland do, that puts more pressure on your defence because you know if you concede then you are in big trouble.\nAnd the Black Cats have problems at the back as well - their only clean sheet in 12 matches under Moyes was against League One side Shrewsbury Town in the EFL Cup.\nIt does not bode well against an Arsenal side that are averaging more than two goals a game this season.\nIt is hard to find any positives from Sunderland's situation but at least they have not been cut adrift at the bottom - yet.\nUnless they win soon, that could happen. I think Hull are also in for a very tough season but when I look at the other two teams immediately above them, Boro and Swansea, they definitely have more about them than the Black Cats do.\nMedia playback is not supported on this device\nChanging manager has clearly not helped Sunderland and comparisons with his predecessor do not help Moyes much either.\nYou cannot tell me that, if Allardyce was still in charge, Sunderland would have only picked up two points so far. It just would not have happened.\nMoyes replaced him relatively late in the summer, which is difficult in itself, but he can only complain about the things that have gone against him up to a point. He should be doing much better than he is.\nHe is still the manager and he is capable of turning things around, so it is right there is no suggestion of him getting the sack.\nBut that will not last forever. This industry is results-driven and Moyes' results are not good enough.\nThat clearly has to change soon and, looking at Sunderland's next few fixtures, the one that stands out as a must-win is their home game against Hull on 19 November.\nIf they fail to beat Arsenal and Bournemouth, then the visit of the Tigers will be the game to define Moyes' tenure. If Sunderland are still without a win after that, things will become extremely difficult for him.\nChris Sutton was speaking to BBC Sport's Chris Bevan.", "summary": "We are exactly a quarter of the way through the Premier League season and some teams at the bottom of the table seem to be turning things around after making a bad start."}
{"document": "The win keeps the Candystripes two points behind leaders Dundalk who won 2-0 away to Shamrock Rovers.\nFormer Plymouth striker Patterson scored his sixth goal of the season in the 14th minute at the Brandywell.\nHe shot into an empty net after the ball broke to him when keeper Dean Delany thwarted Barry McNamee.\nKurtis Byrne should have netted a speedy equaliser but the son of former Celtic player Paul Byrne completely missed his kick in front of goal.\nThat was the one big scare for Kenny Shiels' men on a night when both keepers had a quiet night.\nDerry City have won six and drawn two in the eight games they have played since losing to Finn Harps on the first day of the season.", "summary": "Rory Patterson's early goal proved enough to give second-placed Derry City a home victory over Bohemians in Friday night's Premier Division clash."}
{"document": "The centre-right coalition led by Mr Passos Coelho won the most seats in the election on 4 October.\nBut Socialist leader Antonio Costa has been working to build a coalition with far-left parties.\nMany believe that Mr Passos Coelho will fail to pass the test of a vote of no confidence in Portugal's parliament.\nPresident Anibal Cavaco Silva would then be expected to ask the left to form a government.\nThere are fears that weeks of uncertainty could harm Portugal's economic recovery, more than a year after it exited the strict terms of its รขโยฌ78bn (รยฃ57bn) international bailout.\nEU officials have threatened to take action against Portugal for missing a 15 October deadline to present its draft 2016 budget.\nPortugal is still running one of the highest budget deficits in the eurozone.\n12%\nof the workforce is unemployed\n20%\nof people live below the poverty line\n485,000 emigrated from Portugal between 2011 and 2014\n125% debt to GDP - the second highest rate in the European Union\nMr Passos Coelho's Social Democrats have promised to present a budget, but the two left-wing parties campaigned strongly against his outgoing government's record of harsh austerity.\nThe Left Bloc is seen as allied to the anti-austerity Syriza party in Greece, which for months tried to renegotiate the terms of Greece's eurozone bailout.\nPortugal's Communist Party is regarded as anti-euro and anti-Nato, although it is thought to have moderated its eurozone policies in recent weeks.\nIf Mr Costa's Socialists are eventually chosen to lead a left-wing coalition, it would be the first time since the fall of Portugal's dictatorship in 1974 that a right-wing president appointed a government backed by communists.\nAfter his re-appointment as prime minister leading a right-of-centre coalition, Pedro Passos Coelho has 10 days to appoint ministers and secure parliamentary approval.\nThat may prove impossible, since his coalition lost its majority in the 4 October election and the Socialists have pledged to reject his programme if their talks with other parties succeed.\nTogether, the Socialists, Left Bloc and Communist Party have a majority. All wanted the president to appoint Mr Costa - arguing that anything else was a waste of time.\nIf Mr Passos Coelho does fail, the president could then appoint Mr Costa or keep the incumbent on as caretaker.\nFresh legislative elections may only take place from June, after voters have elected a new president early next year.", "summary": "The Portuguese president has invited incumbent Prime Minister Pedro Passos Coelho to form the next government, despite him having lost his majority."}
{"document": "Nev Edwards scored an early try for Sale, before Castres' Florian Vialelle went over, but Julien Dumora's penalty put the hosts 10-7 ahead at the break.\nJoe Ford sent over a penalty before Castres' Marc-Antoine Rallier and Sales' Will Addison were sin-binned.\nJulien Caminati's late attempt to stop Charlie Ingall saw Sale awarded the decisive penalty try.\nThe win moves the English Premiership side to within one point of Pool Two leaders Newport Gwent Dragons after three games.\nSale got off to the ideal start, Edwards sprinting away for the game's opening points from an Andrei Ostrikov kick, but Castres heaped the pressure on in search of a reply, which came through Vialelle on eight minutes.\nSharks flanker Magnus Lund was forced off with a head injury before the television match official denied Castres a second try, with replays showing that the Sharks defence did enough to force full-back Caminati into touch.\nFord had a chance to put Sale ahead again, but his penalty on 27 minutes drifted wide. Dumora, however, made no mistake soon after, slotting over to give the French side the lead on 33 minutes.\nA combination of probing grubber kicks and scrappy play eventually led to Ford teeing up his second penalty attempt, with the fly-half this time booting the three points to make it 10-10.\nRallier's yellow card following a scuffle saw Ford opt for the posts soon after, but he was off target again before Sales' one-man advantage was lost as Addison was sin-binned.\nSharks pushed for the breakthrough as Ingall went close to touching down, and the video referee eventually gave the penalty try after deciding that Caminati's attempt to stop the winger was illegal.\nCastres: Caminati; Martial, Vialelle, Combezou, Decrop; Dumora, Dupont; Taumoepeau, Rallier, Montes; Samson, Moreaux, Caballero, Diarra, Beattie.\nReplacements: Beziat, Tichit, Martinez, Desroche, Babillot, Fontaine, Lamerat, Seron.\nSale: Arscott; Edwards, Addison, Jennings, Ingall; Ford, Mitchell, Lewis-Roberts, Briggs, Mujati, Mills, Ostrikov, Lund, Seymour (capt), Easter.\nReplacements: Taylor, Flynn, Parker, Beaumont, Neild, Jeffers, James, Haley.\nReferee: David Wilkinson (Ireland)", "summary": "A late penalty try gave Sale victory over Castres at Stade Pierre-Antoine in their European Challenge Cup clash."}
{"document": "The 33-year-old was released by Norwich this summer after five years at the club, during which time he made 75 Canaries first-team appearances.\nTurner also had spells on loan at Fulham and Sheffield Wednesday during his time at Carrow Road.\nIn total, the centre-back has made 436 senior career appearances for eight different clubs.\nFind all the latest football transfers on our dedicated page.", "summary": "League One side Southend United have signed former Hull and Norwich defender Michael Turner on a one-year deal."}
{"document": "United contacted St Johnstone this week with a view to speaking to 52-year-old Wright about the job but this approach was rejected by the Saints board.\nThe Tannadice club - bottom of the Premiership - are seeking to replace Jackie McNamara, who left last month.\nDave Bowman took the first team for Saturday's loss to Partick Thistle.\nThe Tangerines have won only once this season and prop up the table with five points from 10 games.\nFormer Northern Ireland goalkeeper Wright, who replaced Steve Lomas at McDiarmid Park in 2013, led St Johnstone to Scottish Cup success in his first season in charge.\nHe has also secured two successive top-six finishes for the Perth side and previously managed in his homeland.", "summary": "St Johnstone boss Tommy Wright is no longer under consideration for the Dundee United manager's job, BBC Scotland has learned."}
{"document": "Media playback is unsupported on your device\n2 November 2014 Last updated at 17:20 GMT\nHomes and businesses were damaged in the storm, but weather experts were not able to confirm it was a tornado.\nNavtej Johal reports.", "summary": "Residents in Coalville in Leicestershire are cleaning up after high winds hit the town."}
{"document": "5 August 2015 Last updated at 06:36 BST\nShe's now 84 and has been telling Newsround the inspiring story of her life before and after that devastating and world-changing event.\nThis animation contains some sad moments that you might find upsetting.\nYou can find out more about what happened in Hiroshima here.\nWatch 'Hiroshima: A Newsround Special' - Thursday 6 August at 5.30pm on the CBBC channel and on the Newsround website.", "summary": "Bun Hashizume was 14 years old and lived in Hiroshima, in Japan, when a nuclear bomb was dropped on the city 70 years ago, at the end of World War Two."}
{"document": "But what has been your moment of the year?\nFrom Ben Stokes' 258 off 198 balls against South Africa to Stuart Broad's 6-17 against the same opponents, and Alastair Cook being the first Englishman to reach 10,000 Test runs, there are lots of highlights.\nOr perhaps you revelled in Australia being skittled for just 85? Or the dog that invaded the pitch at Vizag?\nThe cricket brains of BBC Sport and BBC Radio 5 live asked you to rank your top 10, and your shortlist will be revealed on Tuesday's Tuffers and Vaughan Cricket Show (20:30 GMT, BBC Radio 5 live and online).\nVotes will no longer count but you can still pick your top 10 and share with friends.\nWhat are your top 10 cricketing moments from this year?", "summary": "It's been topsy-turvy for the England side but eventful and entertaining nonetheless."}
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/conll/sample.json
|
{"words": ["He", "was", "the", "27th", "pitcher", "used", "by", "the", "Angels", "this", "season", ",", "tying", "a", "major-league", "record", "."], "ner": ["O", "O", "O", "O", "O", "O", "O", "O", "B-ORG", "O", "O", "O", "O", "O", "O", "O", "O"]}
{"words": ["CHICAGO", "AT", "ATLANTA"], "ner": ["B-ORG", "O", "B-LOC"]}
{"words": ["President", "Bill", "Clinton", "earlier", "this", "month", "invoked", "special", "powers", "to", "appoint", "Fowler", "during", "the", "congressional", "recess", "because", "the", "Senate", "delayed", "confirming", "his", "nomination", "."], "ner": ["O", "B-PER", "I-PER", "O", "O", "O", "O", "O", "O", "O", "O", "B-PER", "O", "O", "O", "O", "O", "O", "B-ORG", "O", "O", "O", "O", "O"]}
{"words": ["goals", "for", ",", "goals", "against", ",", "points", ")", "."], "ner": ["O", "O", "O", "O", "O", "O", "O", "O", "O"]}
{"words": ["\"", "It", "is", "one", "step", "short", "of", "an", "emergency", "situation", ",", "\"", "a", "police", "spokesman", "said", "via", "telephone", "from", "a", "command", "post", "in", "the", "bush", "."], "ner": ["O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O"]}
{"words": ["U.S.", "Ambassador", "Myles", "Frechette", "applauded", "the", "move", ",", "saying", "it", "could", "prompt", "the", "Clinton", "administration", "to", "remove", "Colombia", "from", "a", "list", "of", "outcast", "nations", "that", "have", "failed", "to", "cooperate", "in", "U.S.", "counternarcotics", "efforts", "."], "ner": ["B-LOC", "O", "B-PER", "I-PER", "O", "O", "O", "O", "O", "O", "O", "O", "O", "B-PER", "O", "O", "O", "B-LOC", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "B-LOC", "O", "O", "O"]}
{"words": ["Halftime"], "ner": ["O"]}
{"words": ["It", "has", "manufacturing", "plants", "in", "San", "Diego", ";", "Creedmoor", ",", "N.C.", ";", "Hampshire", ",", "England", ";", "and", "Tijuana", ",", "Mexico", ",", "and", "distributes", "its", "prodcuts", "in", "more", "than", "120", "countries", "."], "ner": ["O", "O", "O", "O", "O", "B-LOC", "I-LOC", "O", "B-LOC", "O", "B-LOC", "O", "B-LOC", "O", "B-LOC", "O", "O", "B-LOC", "O", "B-LOC", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O"]}
{"words": ["Scotland", "manager", "Craig", "Brown", "said", "on", "Thursday", ":", "\"", "I", "'ve", "watched", "Duncan", "Ferguson", "in", "action", "twice", "recently", "and", "he", "'s", "bang", "in", "form", "."], "ner": ["B-LOC", "O", "B-PER", "I-PER", "O", "O", "O", "O", "O", "O", "O", "O", "B-PER", "I-PER", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O"]}
{"words": ["Clinton", "flew", "in", "by", "helicopter", "from", "Michigan", "City", ",", "Indiana", ",", "after", "ending", "a", "four-day", ",", "559-mile", "trip", "aboard", "a", "campaign", "train", "from", "Washington", "."], "ner": ["B-PER", "O", "O", "O", "O", "O", "B-LOC", "I-LOC", "O", "B-LOC", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "O", "B-LOC", "O"]}
| 0
|
hf_public_repos/transformers/tests/fixtures/tests_samples
|
hf_public_repos/transformers/tests/fixtures/tests_samples/wmt16/sample.json
|
{"translation": {"en": "Membership of Parliament: see Minutes", "ro": "Componenลฃa Parlamentului: a se vedea procesul-verbal"}}
{"translation": {"en": "Approval of Minutes of previous sitting: see Minutes", "ro": "Aprobarea procesului-verbal al ลedinลฃei precedente: a se vedea procesul-verbal"}}
{"translation": {"en": "Membership of Parliament: see Minutes", "ro": "Componenลฃa Parlamentului: a se vedea procesul-verbal"}}
{"translation": {"en": "Verification of credentials: see Minutes", "ro": "Verificarea prerogativelor: a se vedea procesul-verbal"}}
{"translation": {"en": "Documents received: see Minutes", "ro": "Depunere de documente: a se vedea procesul-verbal"}}
{"translation": {"en": "Written statements and oral questions (tabling): see Minutes", "ro": "Declaraลฃii scrise ลi รฎntrebฤri orale (depunere): consultaลฃi procesul-verbal"}}
{"translation": {"en": "Petitions: see Minutes", "ro": "Petiลฃii: a se vedea procesul-verbal"}}
{"translation": {"en": "Texts of agreements forwarded by the Council: see Minutes", "ro": "Transmiterea de cฤtre Consiliu a textelor acordurilor: a se vedea procesul-verbal"}}
{"translation": {"en": "Action taken on Parliament's resolutions: see Minutes", "ro": "Cursul dat rezoluลฃiilor Parlamentului: a se vedea procesul-verbal"}}
{"translation": {"en": "Agenda for next sitting: see Minutes", "ro": "Ordinea de zi a urmฤtoarei ลedinลฃe: a se vedea procesul-verbal"}}
| 0
|
hf_public_repos/transformers/tests
|
hf_public_repos/transformers/tests/optimization/test_optimization_tf.py
|
# Copyright 2020 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import unittest
from transformers import is_tf_available
from transformers.testing_utils import require_tf
if is_tf_available():
import tensorflow as tf
from tensorflow.python.eager import context
from tensorflow.python.framework import ops
from transformers import GradientAccumulator, create_optimizer
@require_tf
class OptimizationFTest(unittest.TestCase):
def assertListAlmostEqual(self, list1, list2, tol):
self.assertEqual(len(list1), len(list2))
for a, b in zip(list1, list2):
self.assertAlmostEqual(a, b, delta=tol)
def testGradientAccumulator(self):
accumulator = GradientAccumulator()
accumulator([tf.constant([1.0, 2.0])])
accumulator([tf.constant([-2.0, 1.0])])
accumulator([tf.constant([-1.0, 2.0])])
with self.assertRaises(ValueError):
accumulator([tf.constant([1.0, 1.0]), tf.constant([2.0, 2.0])])
self.assertEqual(accumulator.step, 3)
self.assertEqual(len(accumulator.gradients), 1)
self.assertListAlmostEqual(accumulator.gradients[0].numpy().tolist(), [-2.0, 5.0], tol=1e-2)
accumulator.reset()
self.assertEqual(accumulator.step, 0)
self.assertListAlmostEqual(accumulator.gradients[0].numpy().tolist(), [0.0, 0.0], tol=1e-2)
def testGradientAccumulatorDistributionStrategy(self):
context._context = None
ops.enable_eager_execution_internal()
physical_devices = tf.config.list_physical_devices("CPU")
if len(physical_devices) == 1:
tf.config.set_logical_device_configuration(
physical_devices[0], [tf.config.LogicalDeviceConfiguration(), tf.config.LogicalDeviceConfiguration()]
)
devices = tf.config.list_logical_devices(device_type="CPU")
strategy = tf.distribute.MirroredStrategy(devices=devices[:2])
with strategy.scope():
accumulator = GradientAccumulator()
variable = tf.Variable([4.0, 3.0])
optimizer, _ = create_optimizer(5e-5, 10, 5)
gradient_placeholder = tf.Variable([0.0, 0.0], trainable=False)
def accumulate_on_replica(gradient):
accumulator([gradient])
def apply_on_replica():
optimizer.apply_gradients(list(zip(accumulator.gradients, [variable])))
@tf.function
def accumulate(grad1, grad2):
with strategy.scope():
local_variables = strategy.experimental_local_results(gradient_placeholder)
local_variables[0].assign(grad1)
local_variables[1].assign(grad2)
strategy.run(accumulate_on_replica, args=(gradient_placeholder,))
@tf.function
def apply_grad():
with strategy.scope():
strategy.run(apply_on_replica)
def _check_local_values(grad1, grad2):
values = strategy.experimental_local_results(accumulator._gradients[0])
self.assertListAlmostEqual(values[0].value(), grad1, tol=1e-2)
self.assertListAlmostEqual(values[1].value(), grad2, tol=1e-2)
accumulate([1.0, 2.0], [-1.0, 1.0])
accumulate([3.0, -1.0], [-1.0, -1.0])
accumulate([-2.0, 2.0], [3.0, -2.0])
self.assertEqual(accumulator.step, 3)
_check_local_values([2.0, 3.0], [1.0, -2.0])
apply_grad()
self.assertListAlmostEqual(variable.value(), [4.0, 3.0], tol=1e-2)
accumulator.reset()
self.assertEqual(accumulator.step, 0)
_check_local_values([0.0, 0.0], [0.0, 0.0])
| 0
|
hf_public_repos/transformers/tests
|
hf_public_repos/transformers/tests/optimization/test_optimization.py
|
# coding=utf-8
# Copyright 2020 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import os
import tempfile
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_torch
if is_torch_available():
import torch
from torch import nn
from transformers import (
Adafactor,
AdamW,
get_constant_schedule,
get_constant_schedule_with_warmup,
get_cosine_schedule_with_warmup,
get_cosine_with_hard_restarts_schedule_with_warmup,
get_inverse_sqrt_schedule,
get_linear_schedule_with_warmup,
get_polynomial_decay_schedule_with_warmup,
)
def unwrap_schedule(scheduler, num_steps=10):
lrs = []
for _ in range(num_steps):
lrs.append(scheduler.get_lr()[0])
scheduler.step()
return lrs
def unwrap_and_save_reload_schedule(scheduler, num_steps=10):
lrs = []
for step in range(num_steps):
lrs.append(scheduler.get_lr()[0])
scheduler.step()
if step == num_steps // 2:
with tempfile.TemporaryDirectory() as tmpdirname:
file_name = os.path.join(tmpdirname, "schedule.bin")
torch.save(scheduler.state_dict(), file_name)
state_dict = torch.load(file_name)
scheduler.load_state_dict(state_dict)
return lrs
@require_torch
class OptimizationTest(unittest.TestCase):
def assertListAlmostEqual(self, list1, list2, tol):
self.assertEqual(len(list1), len(list2))
for a, b in zip(list1, list2):
self.assertAlmostEqual(a, b, delta=tol)
def test_adam_w(self):
w = torch.tensor([0.1, -0.2, -0.1], requires_grad=True)
target = torch.tensor([0.4, 0.2, -0.5])
criterion = nn.MSELoss()
# No warmup, constant schedule, no gradient clipping
optimizer = AdamW(params=[w], lr=2e-1, weight_decay=0.0)
for _ in range(100):
loss = criterion(w, target)
loss.backward()
optimizer.step()
w.grad.detach_() # No zero_grad() function on simple tensors. we do it ourselves.
w.grad.zero_()
self.assertListAlmostEqual(w.tolist(), [0.4, 0.2, -0.5], tol=1e-2)
def test_adafactor(self):
w = torch.tensor([0.1, -0.2, -0.1], requires_grad=True)
target = torch.tensor([0.4, 0.2, -0.5])
criterion = nn.MSELoss()
# No warmup, constant schedule, no gradient clipping
optimizer = Adafactor(
params=[w],
lr=1e-2,
eps=(1e-30, 1e-3),
clip_threshold=1.0,
decay_rate=-0.8,
beta1=None,
weight_decay=0.0,
relative_step=False,
scale_parameter=False,
warmup_init=False,
)
for _ in range(1000):
loss = criterion(w, target)
loss.backward()
optimizer.step()
w.grad.detach_() # No zero_grad() function on simple tensors. we do it ourselves.
w.grad.zero_()
self.assertListAlmostEqual(w.tolist(), [0.4, 0.2, -0.5], tol=1e-2)
@require_torch
class ScheduleInitTest(unittest.TestCase):
m = nn.Linear(50, 50) if is_torch_available() else None
optimizer = AdamW(m.parameters(), lr=10.0) if is_torch_available() else None
num_steps = 10
def assertListAlmostEqual(self, list1, list2, tol, msg=None):
self.assertEqual(len(list1), len(list2))
for a, b in zip(list1, list2):
self.assertAlmostEqual(a, b, delta=tol, msg=msg)
def test_schedulers(self):
common_kwargs = {"num_warmup_steps": 2, "num_training_steps": 10}
# schedulers doct format
# function: (sched_args_dict, expected_learning_rates)
scheds = {
get_constant_schedule: ({}, [10.0] * self.num_steps),
get_constant_schedule_with_warmup: (
{"num_warmup_steps": 4},
[0.0, 2.5, 5.0, 7.5, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0],
),
get_linear_schedule_with_warmup: (
{**common_kwargs},
[0.0, 5.0, 10.0, 8.75, 7.5, 6.25, 5.0, 3.75, 2.5, 1.25],
),
get_cosine_schedule_with_warmup: (
{**common_kwargs},
[0.0, 5.0, 10.0, 9.61, 8.53, 6.91, 5.0, 3.08, 1.46, 0.38],
),
get_cosine_with_hard_restarts_schedule_with_warmup: (
{**common_kwargs, "num_cycles": 2},
[0.0, 5.0, 10.0, 8.53, 5.0, 1.46, 10.0, 8.53, 5.0, 1.46],
),
get_polynomial_decay_schedule_with_warmup: (
{**common_kwargs, "power": 2.0, "lr_end": 1e-7},
[0.0, 5.0, 10.0, 7.656, 5.625, 3.906, 2.5, 1.406, 0.625, 0.156],
),
get_inverse_sqrt_schedule: (
{"num_warmup_steps": 2},
[0.0, 5.0, 10.0, 8.165, 7.071, 6.325, 5.774, 5.345, 5.0, 4.714],
),
}
for scheduler_func, data in scheds.items():
kwargs, expected_learning_rates = data
scheduler = scheduler_func(self.optimizer, **kwargs)
self.assertEqual(len([scheduler.get_lr()[0]]), 1)
lrs_1 = unwrap_schedule(scheduler, self.num_steps)
self.assertListAlmostEqual(
lrs_1,
expected_learning_rates,
tol=1e-2,
msg=f"failed for {scheduler_func} in normal scheduler",
)
scheduler = scheduler_func(self.optimizer, **kwargs)
if scheduler_func.__name__ != "get_constant_schedule":
LambdaScheduleWrapper.wrap_scheduler(scheduler) # wrap to test picklability of the schedule
lrs_2 = unwrap_and_save_reload_schedule(scheduler, self.num_steps)
self.assertListEqual(lrs_1, lrs_2, msg=f"failed for {scheduler_func} in save and reload")
class LambdaScheduleWrapper:
"""See https://github.com/huggingface/transformers/issues/21689"""
def __init__(self, fn):
self.fn = fn
def __call__(self, *args, **kwargs):
return self.fn(*args, **kwargs)
@classmethod
def wrap_scheduler(self, scheduler):
scheduler.lr_lambdas = list(map(self, scheduler.lr_lambdas))
| 0
|
hf_public_repos/transformers
|
hf_public_repos/transformers/notebooks/README.md
|
<!---
Copyright 2023 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->
# ๐ค Transformers Notebooks
You can find here a list of the official notebooks provided by Hugging Face.
Also, we would like to list here interesting content created by the community.
If you wrote some notebook(s) leveraging ๐ค Transformers and would like to be listed here, please open a
Pull Request so it can be included under the Community notebooks.
## Hugging Face's notebooks ๐ค
### Documentation notebooks
You can open any page of the documentation as a notebook in Colab (there is a button directly on said pages) but they are also listed here if you need them:
| Notebook | Description | | |
|:----------|:-------------|:-------------|------:|
| [Quicktour of the library](https://github.com/huggingface/notebooks/blob/main/transformers_doc/en/quicktour.ipynb) | A presentation of the various APIs in Transformers |[](https://colab.research.google.com/github/huggingface/notebooks/blob/main/transformers_doc/en/quicktour.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/en/transformers_doc/quicktour.ipynb)|
| [Summary of the tasks](https://github.com/huggingface/notebooks/blob/main/transformers_doc/en/task_summary.ipynb) | How to run the models of the Transformers library task by task |[](https://colab.research.google.com/github/huggingface/notebooks/blob/main/transformers_doc/en/task_summary.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/transformers_doc/en/task_summary.ipynb)|
| [Preprocessing data](https://github.com/huggingface/notebooks/blob/main/transformers_doc/en/preprocessing.ipynb) | How to use a tokenizer to preprocess your data |[](https://colab.research.google.com/github/huggingface/notebooks/blob/main/transformers_doc/en/preprocessing.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/transformers_doc/en/preprocessing.ipynb)|
| [Fine-tuning a pretrained model](https://github.com/huggingface/notebooks/blob/main/transformers_doc/en/training.ipynb) | How to use the Trainer to fine-tune a pretrained model |[](https://colab.research.google.com/github/huggingface/notebooks/blob/main/transformers_doc/en/training.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/transformers_doc/en/training.ipynb)|
| [Summary of the tokenizers](https://github.com/huggingface/notebooks/blob/main/transformers_doc/en/tokenizer_summary.ipynb) | The differences between the tokenizers algorithm |[](https://colab.research.google.com/github/huggingface/notebooks/blob/main/transformers_doc/en/tokenizer_summary.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/transformers_doc/en/tokenizer_summary.ipynb)|
| [Multilingual models](https://github.com/huggingface/notebooks/blob/main/transformers_doc/en/multilingual.ipynb) | How to use the multilingual models of the library |[](https://colab.research.google.com/github/huggingface/notebooks/blob/main/transformers_doc/en/multilingual.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/transformers_doc/en/multilingual.ipynb)|
### PyTorch Examples
#### Natural Language Processing[[pytorch-nlp]]
| Notebook | Description | | |
|:----------|:-------------|:-------------|------:|
| [Train your tokenizer](https://github.com/huggingface/notebooks/blob/main/examples/tokenizer_training.ipynb) | How to train and use your very own tokenizer |[](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/tokenizer_training.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/tokenizer_training.ipynb)|
| [Train your language model](https://github.com/huggingface/notebooks/blob/main/examples/language_modeling_from_scratch.ipynb) | How to easily start using transformers |[](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/language_modeling_from_scratch.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/language_modeling_from_scratch.ipynb)|
| [How to fine-tune a model on text classification](https://github.com/huggingface/notebooks/blob/main/examples/text_classification.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on any GLUE task. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/text_classification.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/text_classification.ipynb)|
| [How to fine-tune a model on language modeling](https://github.com/huggingface/notebooks/blob/main/examples/language_modeling.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on a causal or masked LM task. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/language_modeling.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/language_modeling.ipynb)|
| [How to fine-tune a model on token classification](https://github.com/huggingface/notebooks/blob/main/examples/token_classification.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on a token classification task (NER, PoS). | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/token_classification.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/token_classification.ipynb)|
| [How to fine-tune a model on question answering](https://github.com/huggingface/notebooks/blob/main/examples/question_answering.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on SQUAD. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/question_answering.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/question_answering.ipynb)|
| [How to fine-tune a model on multiple choice](https://github.com/huggingface/notebooks/blob/main/examples/multiple_choice.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on SWAG. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/multiple_choice.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/multiple_choice.ipynb)|
| [How to fine-tune a model on translation](https://github.com/huggingface/notebooks/blob/main/examples/translation.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on WMT. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/translation.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/translation.ipynb)|
| [How to fine-tune a model on summarization](https://github.com/huggingface/notebooks/blob/main/examples/summarization.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on XSUM. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/summarization.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/summarization.ipynb)|
| [How to train a language model from scratch](https://github.com/huggingface/blog/blob/main/notebooks/01_how_to_train.ipynb)| Highlight all the steps to effectively train Transformer model on custom data | [](https://colab.research.google.com/github/huggingface/blog/blob/main/notebooks/01_how_to_train.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/blog/blob/main/notebooks/01_how_to_train.ipynb)|
| [How to generate text](https://github.com/huggingface/blog/blob/main/notebooks/02_how_to_generate.ipynb)| How to use different decoding methods for language generation with transformers | [](https://colab.research.google.com/github/huggingface/blog/blob/main/notebooks/02_how_to_generate.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/blog/blob/main/notebooks/02_how_to_generate.ipynb)|
| [How to generate text (with constraints)](https://github.com/huggingface/blog/blob/main/notebooks/53_constrained_beam_search.ipynb)| How to guide language generation with user-provided constraints | [](https://colab.research.google.com/github/huggingface/blog/blob/main/notebooks/53_constrained_beam_search.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/blog/blob/main/notebooks/53_constrained_beam_search.ipynb)|
| [Reformer](https://github.com/huggingface/blog/blob/main/notebooks/03_reformer.ipynb)| How Reformer pushes the limits of language modeling | [](https://colab.research.google.com/github/patrickvonplaten/blog/blob/main/notebooks/03_reformer.ipynb)| [](https://studiolab.sagemaker.aws/import/github/patrickvonplaten/blog/blob/main/notebooks/03_reformer.ipynb)|
#### Computer Vision[[pytorch-cv]]
| Notebook | Description | | |
|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|------:|
| [How to fine-tune a model on image classification (Torchvision)](https://github.com/huggingface/notebooks/blob/main/examples/image_classification.ipynb) | Show how to preprocess the data using Torchvision and fine-tune any pretrained Vision model on Image Classification | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/image_classification.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/image_classification.ipynb)|
| [How to fine-tune a model on image classification (Albumentations)](https://github.com/huggingface/notebooks/blob/main/examples/image_classification_albumentations.ipynb) | Show how to preprocess the data using Albumentations and fine-tune any pretrained Vision model on Image Classification | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/image_classification_albumentations.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/image_classification_albumentations.ipynb)|
| [How to fine-tune a model on image classification (Kornia)](https://github.com/huggingface/notebooks/blob/main/examples/image_classification_kornia.ipynb) | Show how to preprocess the data using Kornia and fine-tune any pretrained Vision model on Image Classification | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/image_classification_kornia.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/image_classification_kornia.ipynb)|
| [How to perform zero-shot object detection with OWL-ViT](https://github.com/huggingface/notebooks/blob/main/examples/zeroshot_object_detection_with_owlvit.ipynb) | Show how to perform zero-shot object detection on images with text queries | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/zeroshot_object_detection_with_owlvit.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/zeroshot_object_detection_with_owlvit.ipynb)|
| [How to fine-tune an image captioning model](https://github.com/huggingface/notebooks/blob/main/examples/image_captioning_blip.ipynb) | Show how to fine-tune BLIP for image captioning on a custom dataset | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/image_captioning_blip.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/image_captioning_blip.ipynb)|
| [How to build an image similarity system with Transformers](https://github.com/huggingface/notebooks/blob/main/examples/image_similarity.ipynb) | Show how to build an image similarity system | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/image_similarity.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/image_similarity.ipynb)|
| [How to fine-tune a SegFormer model on semantic segmentation](https://github.com/huggingface/notebooks/blob/main/examples/semantic_segmentation.ipynb) | Show how to preprocess the data and fine-tune a pretrained SegFormer model on Semantic Segmentation | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/semantic_segmentation.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/semantic_segmentation.ipynb)|
| [How to fine-tune a VideoMAE model on video classification](https://github.com/huggingface/notebooks/blob/main/examples/video_classification.ipynb) | Show how to preprocess the data and fine-tune a pretrained VideoMAE model on Video Classification | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/video_classification.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/video_classification.ipynb)|
#### Audio[[pytorch-audio]]
| Notebook | Description | | |
|:----------|:-------------|:-------------|------:|
| [How to fine-tune a speech recognition model in English](https://github.com/huggingface/notebooks/blob/main/examples/speech_recognition.ipynb)| Show how to preprocess the data and fine-tune a pretrained Speech model on TIMIT | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/speech_recognition.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/speech_recognition.ipynb)|
| [How to fine-tune a speech recognition model in any language](https://github.com/huggingface/notebooks/blob/main/examples/multi_lingual_speech_recognition.ipynb)| Show how to preprocess the data and fine-tune a multi-lingually pretrained speech model on Common Voice | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/multi_lingual_speech_recognition.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/multi_lingual_speech_recognition.ipynb)|
| [How to fine-tune a model on audio classification](https://github.com/huggingface/notebooks/blob/main/examples/audio_classification.ipynb)| Show how to preprocess the data and fine-tune a pretrained Speech model on Keyword Spotting | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/audio_classification.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/audio_classification.ipynb)|
#### Biological Sequences[[pytorch-bio]]
| Notebook | Description | | |
|:----------|:----------------------------------------------------------------------------------------|:-------------|------:|
| [How to fine-tune a pre-trained protein model](https://github.com/huggingface/notebooks/blob/main/examples/protein_language_modeling.ipynb) | See how to tokenize proteins and fine-tune a large pre-trained protein "language" model | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/protein_language_modeling.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/protein_language_modeling.ipynb) |
| [How to generate protein folds](https://github.com/huggingface/notebooks/blob/main/examples/protein_folding.ipynb) | See how to go from protein sequence to a full protein model and PDB file | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/protein_folding.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/protein_folding.ipynb) |
| [How to fine-tune a Nucleotide Transformer model](https://github.com/huggingface/notebooks/blob/main/examples/nucleotide_transformer_dna_sequence_modelling.ipynb) | See how to tokenize DNA and fine-tune a large pre-trained DNA "language" model | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/nucleotide_transformer_dna_sequence_modelling.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/nucleotide_transformer_dna_sequence_modelling.ipynb) |
| [Fine-tune a Nucleotide Transformer model with LoRA](https://github.com/huggingface/notebooks/blob/main/examples/nucleotide_transformer_dna_sequence_modelling_with_peft.ipynb) | Train even larger DNA models in a memory-efficient way | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/nucleotide_transformer_dna_sequence_modelling_with_peft.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/nucleotide_transformer_dna_sequence_modelling_with_peft.ipynb) |
#### Other modalities[[pytorch-other]]
| Notebook | Description | | |
|:----------|:----------------------------------------------------------------------------------------|:-------------|------:|
| [Probabilistic Time Series Forecasting](https://github.com/huggingface/notebooks/blob/main/examples/time-series-transformers.ipynb) | See how to train Time Series Transformer on a custom dataset | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/time-series-transformers.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/time-series-transformers.ipynb) |
#### Utility notebooks[[pytorch-utility]]
| Notebook | Description | | |
|:----------|:-------------|:-------------|------:|
| [How to export model to ONNX](https://github.com/huggingface/notebooks/blob/main/examples/onnx-export.ipynb)| Highlight how to export and run inference workloads through ONNX |
| [How to use Benchmarks](https://github.com/huggingface/notebooks/blob/main/examples/benchmark.ipynb)| How to benchmark models with transformers | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/benchmark.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/benchmark.ipynb)|
### TensorFlow Examples
#### Natural Language Processing[[tensorflow-nlp]]
| Notebook | Description | | |
|:----------|:-------------|:-------------|------:|
| [Train your tokenizer](https://github.com/huggingface/notebooks/blob/main/examples/tokenizer_training.ipynb) | How to train and use your very own tokenizer |[](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/tokenizer_training.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/tokenizer_training.ipynb)|
| [Train your language model](https://github.com/huggingface/notebooks/blob/main/examples/language_modeling_from_scratch-tf.ipynb) | How to easily start using transformers |[](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/language_modeling_from_scratch-tf.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/language_modeling_from_scratch-tf.ipynb)|
| [How to fine-tune a model on text classification](https://github.com/huggingface/notebooks/blob/main/examples/text_classification-tf.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on any GLUE task. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/text_classification-tf.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/text_classification-tf.ipynb)|
| [How to fine-tune a model on language modeling](https://github.com/huggingface/notebooks/blob/main/examples/language_modeling-tf.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on a causal or masked LM task. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/language_modeling-tf.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/language_modeling-tf.ipynb)|
| [How to fine-tune a model on token classification](https://github.com/huggingface/notebooks/blob/main/examples/token_classification-tf.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on a token classification task (NER, PoS). | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/token_classification-tf.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/token_classification-tf.ipynb)|
| [How to fine-tune a model on question answering](https://github.com/huggingface/notebooks/blob/main/examples/question_answering-tf.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on SQUAD. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/question_answering-tf.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/question_answering-tf.ipynb)|
| [How to fine-tune a model on multiple choice](https://github.com/huggingface/notebooks/blob/main/examples/multiple_choice-tf.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on SWAG. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/multiple_choice-tf.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/multiple_choice-tf.ipynb)|
| [How to fine-tune a model on translation](https://github.com/huggingface/notebooks/blob/main/examples/translation-tf.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on WMT. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/translation-tf.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/translation-tf.ipynb)|
| [How to fine-tune a model on summarization](https://github.com/huggingface/notebooks/blob/main/examples/summarization-tf.ipynb)| Show how to preprocess the data and fine-tune a pretrained model on XSUM. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/summarization-tf.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/summarization-tf.ipynb)|
#### Computer Vision[[tensorflow-cv]]
| Notebook | Description | | |
|:---------------------------------------------------------------------------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------------------------|:-------------|------:|
| [How to fine-tune a model on image classification](https://github.com/huggingface/notebooks/blob/main/examples/image_classification-tf.ipynb) | Show how to preprocess the data and fine-tune any pretrained Vision model on Image Classification | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/image_classification-tf.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/image_classification-tf.ipynb)|
| [How to fine-tune a SegFormer model on semantic segmentation](https://github.com/huggingface/notebooks/blob/main/examples/semantic_segmentation-tf.ipynb) | Show how to preprocess the data and fine-tune a pretrained SegFormer model on Semantic Segmentation | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/semantic_segmentation-tf.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/semantic_segmentation-tf.ipynb)|
#### Biological Sequences[[tensorflow-bio]]
| Notebook | Description | | |
|:----------|:-------------|:-------------|------:|
| [How to fine-tune a pre-trained protein model](https://github.com/huggingface/notebooks/blob/main/examples/protein_language_modeling-tf.ipynb) | See how to tokenize proteins and fine-tune a large pre-trained protein "language" model | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/protein_language_modeling-tf.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/protein_language_modeling-tf.ipynb) |
#### Utility notebooks[[tensorflow-utility]]
| Notebook | Description | | |
|:----------|:-------------|:-------------|------:|
| [How to train TF/Keras models on TPU](https://github.com/huggingface/notebooks/blob/main/examples/tpu_training-tf.ipynb) | See how to train at high speed on Google's TPU hardware | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/tpu_training-tf.ipynb) | [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/tpu_training-tf.ipynb) |
### Optimum notebooks
๐ค [Optimum](https://github.com/huggingface/optimum) is an extension of ๐ค Transformers, providing a set of performance optimization tools enabling maximum efficiency to train and run models on targeted hardwares.
| Notebook | Description | | |
|:----------|:-------------|:-------------|------:|
| [How to quantize a model with ONNX Runtime for text classification](https://github.com/huggingface/notebooks/blob/main/examples/text_classification_quantization_ort.ipynb)| Show how to apply static and dynamic quantization on a model using [ONNX Runtime](https://github.com/microsoft/onnxruntime) for any GLUE task. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/text_classification_quantization_ort.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/text_classification_quantization_ort.ipynb)|
| [How to quantize a model with Intel Neural Compressor for text classification](https://github.com/huggingface/notebooks/blob/main/examples/text_classification_quantization_inc.ipynb)| Show how to apply static, dynamic and aware training quantization on a model using [Intel Neural Compressor (INC)](https://github.com/intel/neural-compressor) for any GLUE task. | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/text_classification_quantization_inc.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/text_classification_quantization_inc.ipynb)|
| [How to fine-tune a model on text classification with ONNX Runtime](https://github.com/huggingface/notebooks/blob/main/examples/text_classification_ort.ipynb)| Show how to preprocess the data and fine-tune a model on any GLUE task using [ONNX Runtime](https://github.com/microsoft/onnxruntime). | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/text_classification_ort.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/text_classification_ort.ipynb)|
| [How to fine-tune a model on summarization with ONNX Runtime](https://github.com/huggingface/notebooks/blob/main/examples/summarization_ort.ipynb)| Show how to preprocess the data and fine-tune a model on XSUM using [ONNX Runtime](https://github.com/microsoft/onnxruntime). | [](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/summarization_ort.ipynb)| [](https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/main/examples/summarization_ort.ipynb)|
## Community notebooks:
More notebooks developed by the community are available [here](https://hf.co/docs/transformers/community#community-notebooks).
| 0
|
hf_public_repos/transformers
|
hf_public_repos/transformers/docs/README.md
|
<!---
Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->
# Generating the documentation
To generate the documentation, you first have to build it. Several packages are necessary to build the doc,
you can install them with the following command, at the root of the code repository:
```bash
pip install -e ".[docs]"
```
Then you need to install our special tool that builds the documentation:
```bash
pip install git+https://github.com/huggingface/doc-builder
```
---
**NOTE**
You only need to generate the documentation to inspect it locally (if you're planning changes and want to
check how they look before committing for instance). You don't have to commit the built documentation.
---
## Building the documentation
Once you have setup the `doc-builder` and additional packages, you can generate the documentation by
typing the following command:
```bash
doc-builder build transformers docs/source/en/ --build_dir ~/tmp/test-build
```
You can adapt the `--build_dir` to set any temporary folder that you prefer. This command will create it and generate
the MDX files that will be rendered as the documentation on the main website. You can inspect them in your favorite
Markdown editor.
## Previewing the documentation
To preview the docs, first install the `watchdog` module with:
```bash
pip install watchdog
```
Then run the following command:
```bash
doc-builder preview {package_name} {path_to_docs}
```
For example:
```bash
doc-builder preview transformers docs/source/en/
```
The docs will be viewable at [http://localhost:3000](http://localhost:3000). You can also preview the docs once you have opened a PR. You will see a bot add a comment to a link where the documentation with your changes lives.
---
**NOTE**
The `preview` command only works with existing doc files. When you add a completely new file, you need to update `_toctree.yml` & restart `preview` command (`ctrl-c` to stop it & call `doc-builder preview ...` again).
---
## Adding a new element to the navigation bar
Accepted files are Markdown (.md).
Create a file with its extension and put it in the source directory. You can then link it to the toc-tree by putting
the filename without the extension in the [`_toctree.yml`](https://github.com/huggingface/transformers/blob/main/docs/source/en/_toctree.yml) file.
## Renaming section headers and moving sections
It helps to keep the old links working when renaming the section header and/or moving sections from one document to another. This is because the old links are likely to be used in Issues, Forums, and Social media and it'd make for a much more superior user experience if users reading those months later could still easily navigate to the originally intended information.
Therefore, we simply keep a little map of moved sections at the end of the document where the original section was. The key is to preserve the original anchor.
So if you renamed a section from: "Section A" to "Section B", then you can add at the end of the file:
```
Sections that were moved:
[ <a href="#section-b">Section A</a><a id="section-a"></a> ]
```
and of course, if you moved it to another file, then:
```
Sections that were moved:
[ <a href="../new-file#section-b">Section A</a><a id="section-a"></a> ]
```
Use the relative style to link to the new file so that the versioned docs continue to work.
For an example of a rich moved section set please see the very end of [the Trainer doc](https://github.com/huggingface/transformers/blob/main/docs/source/en/main_classes/trainer.md).
## Writing Documentation - Specification
The `huggingface/transformers` documentation follows the
[Google documentation](https://sphinxcontrib-napoleon.readthedocs.io/en/latest/example_google.html) style for docstrings,
although we can write them directly in Markdown.
### Adding a new tutorial
Adding a new tutorial or section is done in two steps:
- Add a new file under `./source`. This file can either be ReStructuredText (.rst) or Markdown (.md).
- Link that file in `./source/_toctree.yml` on the correct toc-tree.
Make sure to put your new file under the proper section. It's unlikely to go in the first section (*Get Started*), so
depending on the intended targets (beginners, more advanced users, or researchers) it should go in sections two, three, or
four.
### Translating
When translating, refer to the guide at [./TRANSLATING.md](https://github.com/huggingface/transformers/blob/main/docs/TRANSLATING.md).
### Adding a new model
When adding a new model:
- Create a file `xxx.md` or under `./source/model_doc` (don't hesitate to copy an existing file as template).
- Link that file in `./source/_toctree.yml`.
- Write a short overview of the model:
- Overview with paper & authors
- Paper abstract
- Tips and tricks and how to use it best
- Add the classes that should be linked in the model. This generally includes the configuration, the tokenizer, and
every model of that class (the base model, alongside models with additional heads), both in PyTorch and TensorFlow.
The order is generally:
- Configuration
- Tokenizer
- PyTorch base model
- PyTorch head models
- TensorFlow base model
- TensorFlow head models
- Flax base model
- Flax head models
These classes should be added using our Markdown syntax. Usually as follows:
```
## XXXConfig
[[autodoc]] XXXConfig
```
This will include every public method of the configuration that is documented. If for some reason you wish for a method
not to be displayed in the documentation, you can do so by specifying which methods should be in the docs:
```
## XXXTokenizer
[[autodoc]] XXXTokenizer
- build_inputs_with_special_tokens
- get_special_tokens_mask
- create_token_type_ids_from_sequences
- save_vocabulary
```
If you just want to add a method that is not documented (for instance magic methods like `__call__` are not documented
by default) you can put the list of methods to add in a list that contains `all`:
```
## XXXTokenizer
[[autodoc]] XXXTokenizer
- all
- __call__
```
### Writing source documentation
Values that should be put in `code` should either be surrounded by backticks: \`like so\`. Note that argument names
and objects like True, None, or any strings should usually be put in `code`.
When mentioning a class, function, or method, it is recommended to use our syntax for internal links so that our tool
adds a link to its documentation with this syntax: \[\`XXXClass\`\] or \[\`function\`\]. This requires the class or
function to be in the main package.
If you want to create a link to some internal class or function, you need to
provide its path. For instance: \[\`utils.ModelOutput\`\]. This will be converted into a link with
`utils.ModelOutput` in the description. To get rid of the path and only keep the name of the object you are
linking to in the description, add a ~: \[\`~utils.ModelOutput\`\] will generate a link with `ModelOutput` in the description.
The same works for methods so you can either use \[\`XXXClass.method\`\] or \[~\`XXXClass.method\`\].
#### Defining arguments in a method
Arguments should be defined with the `Args:` (or `Arguments:` or `Parameters:`) prefix, followed by a line return and
an indentation. The argument should be followed by its type, with its shape if it is a tensor, a colon, and its
description:
```
Args:
n_layers (`int`): The number of layers of the model.
```
If the description is too long to fit in one line, another indentation is necessary before writing the description
after the argument.
Here's an example showcasing everything so far:
```
Args:
input_ids (`torch.LongTensor` of shape `(batch_size, sequence_length)`):
Indices of input sequence tokens in the vocabulary.
Indices can be obtained using [`AlbertTokenizer`]. See [`~PreTrainedTokenizer.encode`] and
[`~PreTrainedTokenizer.__call__`] for details.
[What are input IDs?](../glossary#input-ids)
```
For optional arguments or arguments with defaults we follow the following syntax: imagine we have a function with the
following signature:
```
def my_function(x: str = None, a: float = 1):
```
then its documentation should look like this:
```
Args:
x (`str`, *optional*):
This argument controls ...
a (`float`, *optional*, defaults to 1):
This argument is used to ...
```
Note that we always omit the "defaults to \`None\`" when None is the default for any argument. Also note that even
if the first line describing your argument type and its default gets long, you can't break it on several lines. You can
however write as many lines as you want in the indented description (see the example above with `input_ids`).
#### Writing a multi-line code block
Multi-line code blocks can be useful for displaying examples. They are done between two lines of three backticks as usual in Markdown:
````
```
# first line of code
# second line
# etc
```
````
We follow the [doctest](https://docs.python.org/3/library/doctest.html) syntax for the examples to automatically test
the results to stay consistent with the library.
#### Writing a return block
The return block should be introduced with the `Returns:` prefix, followed by a line return and an indentation.
The first line should be the type of the return, followed by a line return. No need to indent further for the elements
building the return.
Here's an example of a single value return:
```
Returns:
`List[int]`: A list of integers in the range [0, 1] --- 1 for a special token, 0 for a sequence token.
```
Here's an example of a tuple return, comprising several objects:
```
Returns:
`tuple(torch.FloatTensor)` comprising various elements depending on the configuration ([`BertConfig`]) and inputs:
- ** loss** (*optional*, returned when `masked_lm_labels` is provided) `torch.FloatTensor` of shape `(1,)` --
Total loss is the sum of the masked language modeling loss and the next sequence prediction (classification) loss.
- **prediction_scores** (`torch.FloatTensor` of shape `(batch_size, sequence_length, config.vocab_size)`) --
Prediction scores of the language modeling head (scores for each vocabulary token before SoftMax).
```
#### Adding an image
Due to the rapidly growing repository, it is important to make sure that no files that would significantly weigh down the repository are added. This includes images, videos, and other non-text files. We prefer to leverage a hf.co hosted `dataset` like
the ones hosted on [`hf-internal-testing`](https://huggingface.co/hf-internal-testing) in which to place these files and reference
them by URL. We recommend putting them in the following dataset: [huggingface/documentation-images](https://huggingface.co/datasets/huggingface/documentation-images).
If an external contribution, feel free to add the images to your PR and ask a Hugging Face member to migrate your images
to this dataset.
## Styling the docstring
We have an automatic script running with the `make style` comment that will make sure that:
- the docstrings fully take advantage of the line width
- all code examples are formatted using black, like the code of the Transformers library
This script may have some weird failures if you made a syntax mistake or if you uncover a bug. Therefore, it's
recommended to commit your changes before running `make style`, so you can revert the changes done by that script
easily.
# Testing documentation examples
Good documentation often comes with an example of how a specific function or class should be used.
Each model class should contain at least one example showcasing
how to use this model class in inference. *E.g.* the class [Wav2Vec2ForCTC](https://huggingface.co/docs/transformers/model_doc/wav2vec2#transformers.Wav2Vec2ForCTC)
includes an example of how to transcribe speech to text in the
[docstring of its forward function](https://huggingface.co/docs/transformers/model_doc/wav2vec2#transformers.Wav2Vec2ForCTC.forward).
## Writing documentation examples
The syntax for Example docstrings can look as follows:
```
Example:
```python
>>> from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
>>> from datasets import load_dataset
>>> import torch
>>> dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
>>> dataset = dataset.sort("id")
>>> sampling_rate = dataset.features["audio"].sampling_rate
>>> processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-base-960h")
>>> model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-base-960h")
>>> # 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]
'MISTER QUILTER IS THE APOSTLE OF THE MIDDLE CLASSES AND WE ARE GLAD TO WELCOME HIS GOSPEL'
```
```
The docstring should give a minimal, clear example of how the respective model
is to be used in inference and also include the expected (ideally sensible)
output.
Often, readers will try out the example before even going through the function
or class definitions. Therefore, it is of utmost importance that the example
works as expected.
## Docstring testing
To do so each example should be included in the doctests.
We use pytests' [doctest integration](https://docs.pytest.org/doctest.html) to verify that all of our examples run correctly.
For Transformers, the doctests are run on a daily basis via GitHub Actions as can be
seen [here](https://github.com/huggingface/transformers/actions/workflows/doctests.yml).
### For Python files
Run all the tests in the docstrings of a given file with the following command, here is how we test the modeling file of Wav2Vec2 for instance:
```bash
pytest --doctest-modules src/transformers/models/wav2vec2/modeling_wav2vec2.py -sv --doctest-continue-on-failure
```
If you want to isolate a specific docstring, just add `::` after the file name then type the whole path of the function/class/method whose docstring you want to test. For instance, here is how to just test the forward method of `Wav2Vec2ForCTC`:
```bash
pytest --doctest-modules src/transformers/models/wav2vec2/modeling_wav2vec2.py::transformers.models.wav2vec2.modeling_wav2vec2.Wav2Vec2ForCTC.forward -sv --doctest-continue-on-failure
```
### For Markdown files
You can test locally a given file with this command (here testing the quicktour):
```bash
pytest --doctest-modules docs/source/quicktour.md -sv --doctest-continue-on-failure --doctest-glob="*.md"
```
### Writing doctests
Here are a few tips to help you debug the doctests and make them pass:
- The outputs of the code need to match the expected output **exactly**, so make sure you have the same outputs. In particular doctest will see a difference between single quotes and double quotes, or a missing parenthesis. The only exceptions to that rule are:
* whitespace: one give whitespace (space, tabulation, new line) is equivalent to any number of whitespace, so you can add new lines where there are spaces to make your output more readable.
* numerical values: you should never put more than 4 or 5 digits to expected results as different setups or library versions might get you slightly different results. `doctest` is configured to ignore any difference lower than the precision to which you wrote (so 1e-4 if you write 4 digits).
- Don't leave a block of code that is very long to execute. If you can't make it fast, you can either not use the doctest syntax on it (so that it's ignored), or if you want to use the doctest syntax to show the results, you can add a comment `# doctest: +SKIP` at the end of the lines of code too long to execute
- Each line of code that produces a result needs to have that result written below. You can ignore an output if you don't want to show it in your code example by adding a comment ` # doctest: +IGNORE_RESULT` at the end of the line of code producing it.
| 0
|
hf_public_repos/transformers
|
hf_public_repos/transformers/docs/TRANSLATING.md
|
### Translating the Transformers documentation into your language
As part of our mission to democratize machine learning, we'd love to make the Transformers library available in many more languages! Follow the steps below if you want to help translate the documentation into your language ๐.
**๐๏ธ Open an issue**
To get started, navigate to the [Issues](https://github.com/huggingface/transformers/issues) page of this repo and check if anyone else has opened an issue for your language. If not, open a new issue by selecting the "Translation template" from the "New issue" button.
Once an issue exists, post a comment to indicate which chapters you'd like to work on, and we'll add your name to the list.
**๐ด Fork the repository**
First, you'll need to [fork the Transformers repo](https://docs.github.com/en/get-started/quickstart/fork-a-repo). You can do this by clicking on the **Fork** button on the top-right corner of this repo's page.
Once you've forked the repo, you'll want to get the files on your local machine for editing. You can do that by cloning the fork with Git as follows:
```bash
git clone https://github.com/YOUR-USERNAME/transformers.git
```
**๐ Copy-paste the English version with a new language code**
The documentation files are in one leading directory:
- [`docs/source`](https://github.com/huggingface/transformers/tree/main/docs/source): All the documentation materials are organized here by language.
You'll only need to copy the files in the [`docs/source/en`](https://github.com/huggingface/transformers/tree/main/docs/source/en) directory, so first navigate to your fork of the repo and run the following:
```bash
cd ~/path/to/transformers/docs
cp -r source/en source/LANG-ID
```
Here, `LANG-ID` should be one of the ISO 639-1 or ISO 639-2 language codes -- see [here](https://www.loc.gov/standards/iso639-2/php/code_list.php) for a handy table.
**โ๏ธ Start translating**
The fun part comes - translating the text!
The first thing we recommend is translating the part of the `_toctree.yml` file that corresponds to your doc chapter. This file is used to render the table of contents on the website.
> ๐ If the `_toctree.yml` file doesn't yet exist for your language, you can create one by copy-pasting from the English version and deleting the sections unrelated to your chapter. Just make sure it exists in the `docs/source/LANG-ID/` directory!
The fields you should add are `local` (with the name of the file containing the translation; e.g. `autoclass_tutorial`), and `title` (with the title of the doc in your language; e.g. `Load pretrained instances with an AutoClass`) -- as a reference, here is the `_toctree.yml` for [English](https://github.com/huggingface/transformers/blob/main/docs/source/en/_toctree.yml):
```yaml
- sections:
- local: pipeline_tutorial # Do not change this! Use the same name for your .md file
title: Pipelines for inference # Translate this!
...
title: Tutorials # Translate this!
```
Once you have translated the `_toctree.yml` file, you can start translating the [MDX](https://mdxjs.com/) files associated with your docs chapter.
> ๐ If you'd like others to help you with the translation, you should [open an issue](https://github.com/huggingface/transformers/issues) and tag @stevhliu and @MKhalusova.
| 0
|
hf_public_repos/transformers/docs
|
hf_public_repos/transformers/docs/source/_config.py
|
# docstyle-ignore
INSTALL_CONTENT = """
# Transformers installation
! pip install transformers datasets evaluate
# To install from source instead of the last release, comment the command above and uncomment the following one.
# ! pip install git+https://github.com/huggingface/transformers.git
"""
notebook_first_cells = [{"type": "code", "content": INSTALL_CONTENT}]
black_avoid_patterns = {
"{processor_class}": "FakeProcessorClass",
"{model_class}": "FakeModelClass",
"{object_class}": "FakeObjectClass",
}
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/preprocessing.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Vorverarbeiten
[[open-in-colab]]
Bevor Sie Ihre Daten in einem Modell verwenden kรถnnen, mรผssen die Daten in ein fรผr das Modell akzeptables Format gebracht werden. Ein Modell versteht keine Rohtexte, Bilder oder Audiodaten. Diese Eingaben mรผssen in Zahlen umgewandelt und zu Tensoren zusammengesetzt werden. In dieser Anleitung werden Sie:
* Textdaten mit einem Tokenizer vorverarbeiten.
* Bild- oder Audiodaten mit einem Feature Extractor vorverarbeiten.
* Daten fรผr eine multimodale Aufgabe mit einem Prozessor vorverarbeiten.
## NLP
<Youtube id="Yffk5aydLzg"/>
Das wichtigste Werkzeug zur Verarbeitung von Textdaten ist ein [Tokenizer](main_classes/tokenizer). Ein Tokenizer zerlegt Text zunรคchst nach einer Reihe von Regeln in *Token*. Die Token werden in Zahlen umgewandelt, die zum Aufbau von Tensoren als Eingabe fรผr ein Modell verwendet werden. Alle zusรคtzlichen Eingaben, die ein Modell benรถtigt, werden ebenfalls vom Tokenizer hinzugefรผgt.
<Tip>
Wenn Sie ein vortrainiertes Modell verwenden mรถchten, ist es wichtig, den zugehรถrigen vortrainierten Tokenizer zu verwenden. Dadurch wird sichergestellt, dass der Text auf die gleiche Weise aufgeteilt wird wie das Pretraining-Korpus und die gleichen entsprechenden Token-zu-Index (in der Regel als *vocab* bezeichnet) wรคhrend des Pretrainings verwendet werden.
</Tip>
Laden Sie einen vortrainierten Tokenizer mit der Klasse [AutoTokenizer], um schnell loszulegen. Damit wird das *vocab* heruntergeladen, das verwendet wird, wenn ein Modell vortrainiert wird.
### Tokenize
Laden Sie einen vortrainierten Tokenizer mit [`AutoTokenizer.from_pretrained`]:
```py
>>> from transformers import AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("bert-base-cased")
```
Dann รผbergeben Sie Ihren Satz an den Tokenizer:
```py
>>> encoded_input = tokenizer("Do not meddle in the affairs of wizards, for they are subtle and quick to anger.")
>>> print(encoded_input)
{'input_ids': [101, 2079, 2025, 19960, 10362, 1999, 1996, 3821, 1997, 16657, 1010, 2005, 2027, 2024, 11259, 1998, 4248, 2000, 4963, 1012, 102],
'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}
```
Der Tokenizer gibt ein Wรถrterbuch mit drei wichtigen Elementen zurรผck:
* [input_ids](glossary#input-ids) sind die Indizes, die den einzelnen Token im Satz entsprechen.
* [attention_mask](glossary#attention-mask) gibt an, ob ein Token beachtet werden soll oder nicht.
* [token_type_ids](glossary#token-type-ids) gibt an, zu welcher Sequenz ein Token gehรถrt, wenn es mehr als eine Sequenz gibt.
Sie kรถnnen die `input_ids` dekodieren, um die ursprรผngliche Eingabe zurรผckzugeben:
```py
>>> tokenizer.decode(encoded_input["input_ids"])
'[CLS] Do not meddle in the affairs of wizards, for they are subtle and quick to anger. [SEP]'
```
Wie Sie sehen kรถnnen, hat der Tokenisierer zwei spezielle Token - `CLS` und `SEP` (Klassifikator und Separator) - zum Satz hinzugefรผgt. Nicht alle Modelle benรถtigen
spezielle Token, aber wenn dies der Fall ist, fรผgt der Tokenisierer sie automatisch fรผr Sie hinzu.
Wenn Sie mehrere Sรคtze verarbeiten wollen, รผbergeben Sie die Sรคtze als Liste an den Tokenizer:
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_inputs = tokenizer(batch_sentences)
>>> print(encoded_inputs)
{'input_ids': [[101, 1252, 1184, 1164, 1248, 6462, 136, 102],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102]],
'token_type_ids': [[0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]],
'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1]]}
```
### Pad
Dies bringt uns zu einem wichtigen Thema. Wenn Sie einen Haufen von Sรคtzen verarbeiten, sind diese nicht immer gleich lang. Das ist ein Problem, weil Tensoren, die Eingabe fรผr das Modell, eine einheitliche Form haben mรผssen. Padding ist eine Strategie, die sicherstellt, dass Tensoren rechteckig sind, indem ein spezielles *Padding-Token* zu Sรคtzen mit weniger Token hinzugefรผgt wird.
Setzen Sie den Parameter "padding" auf "true", um die kรผrzeren Sequenzen im Stapel so aufzufรผllen, dass sie der lรคngsten Sequenz entsprechen:
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch_sentences, padding=True)
>>> print(encoded_input)
{'input_ids': [[101, 1252, 1184, 1164, 1248, 6462, 136, 102, 0, 0, 0, 0, 0, 0, 0],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102, 0, 0, 0, 0, 0, 0, 0, 0]],
'token_type_ids': [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]],
'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]]}
```
Beachten Sie, dass der Tokenizer den ersten und den dritten Satz mit einer "0" aufgefรผllt hat, weil sie kรผrzer sind!
### Kรผrzung
Auf der anderen Seite des Spektrums kann es vorkommen, dass eine Sequenz zu lang fรผr ein Modell ist. In diesem Fall mรผssen Sie die Sequenz auf eine kรผrzere Lรคnge kรผrzen.
Setzen Sie den Parameter "truncation" auf "true", um eine Sequenz auf die vom Modell akzeptierte Hรถchstlรคnge zu kรผrzen:
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch_sentences, padding=True, truncation=True)
>>> print(encoded_input)
{'input_ids': [[101, 1252, 1184, 1164, 1248, 6462, 136, 102, 0, 0, 0, 0, 0, 0, 0],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102, 0, 0, 0, 0, 0, 0, 0, 0]],
'token_type_ids': [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]],
'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]]}
```
### Tensoren erstellen
Schlieรlich mรถchten Sie, dass der Tokenizer die tatsรคchlichen Tensoren zurรผckgibt, die dem Modell zugefรผhrt werden.
Setzen Sie den Parameter `return_tensors` entweder auf `pt` fรผr PyTorch, oder `tf` fรผr TensorFlow:
<frameworkcontent>
<pt>
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch_sentences, padding=True, truncation=True, return_tensors="pt")
>>> print(encoded_input)
{'input_ids': tensor([[101, 1252, 1184, 1164, 1248, 6462, 136, 102, 0, 0, 0, 0, 0, 0, 0],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102, 0, 0, 0, 0, 0, 0, 0, 0]]),
'token_type_ids': tensor([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]),
'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]])}
```
</pt>
<tf>
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch_sentences, padding=True, truncation=True, return_tensors="tf")
>>> print(encoded_input)
{'input_ids': <tf.Tensor: shape=(2, 9), dtype=int32, numpy=
array([[101, 1252, 1184, 1164, 1248, 6462, 136, 102, 0, 0, 0, 0, 0, 0, 0],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102, 0, 0, 0, 0, 0, 0, 0, 0]],
dtype=int32)>,
'token_type_ids': <tf.Tensor: shape=(2, 9), dtype=int32, numpy=
array([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], dtype=int32)>,
'attention_mask': <tf.Tensor: shape=(2, 9), dtype=int32, numpy=
array([[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]], dtype=int32)>}
```
</tf>
</frameworkcontent>
## Audio
Audioeingaben werden anders vorverarbeitet als Texteingaben, aber das Endziel bleibt dasselbe: numerische Sequenzen zu erstellen, die das Modell verstehen kann. Ein [feature extractor](main_classes/feature_extractor) dient dem ausdrรผcklichen Zweck, Merkmale aus Rohbild- oder Audiodaten zu extrahieren und in Tensoren zu konvertieren. Bevor Sie beginnen, installieren Sie ๐ค Datasets, um einen Audio-Datensatz zu laden, mit dem Sie experimentieren kรถnnen:
```bash
pip install datasets
```
Laden Sie den [MInDS-14](https://huggingface.co/datasets/PolyAI/minds14) Datensatz (weitere Informationen zum Laden eines Datensatzes finden Sie im ๐ค [Datasets tutorial](https://huggingface.co/docs/datasets/load_hub)):
```py
>>> from datasets import load_dataset, Audio
>>> dataset = load_dataset("PolyAI/minds14", name="en-US", split="train")
```
Greifen Sie auf das erste Element der `audio`-Spalte zu, um einen Blick auf die Eingabe zu werfen. Durch den Aufruf der Spalte "audio" wird die Audiodatei automatisch geladen und neu gesampelt:
```py
>>> dataset[0]["audio"]
{'array': array([ 0. , 0.00024414, -0.00024414, ..., -0.00024414,
0. , 0. ], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/f14948e0e84be638dd7943ac36518a4cf3324e8b7aa331c5ab11541518e9368c/en-US~JOINT_ACCOUNT/602ba55abb1e6d0fbce92065.wav',
'sampling_rate': 8000}
```
Dies gibt drei Elemente zurรผck:
* "array" ist das Sprachsignal, das als 1D-Array geladen - und mรถglicherweise neu gesampelt - wurde.
* Pfad" zeigt auf den Speicherort der Audiodatei.
* `sampling_rate` bezieht sich darauf, wie viele Datenpunkte im Sprachsignal pro Sekunde gemessen werden.
### Resample
Fรผr dieses Tutorial werden Sie das Modell [Wav2Vec2](https://huggingface.co/facebook/wav2vec2-base) verwenden. Wie Sie aus der Modellkarte ersehen kรถnnen, ist das Wav2Vec2-Modell auf 16kHz abgetastetes Sprachaudio vortrainiert. Es ist wichtig, dass die Abtastrate Ihrer Audiodaten mit der Abtastrate des Datensatzes รผbereinstimmt, der fรผr das Pre-Training des Modells verwendet wurde. Wenn die Abtastrate Ihrer Daten nicht dieselbe ist, mรผssen Sie Ihre Audiodaten neu abtasten.
Der Datensatz [MInDS-14](https://huggingface.co/datasets/PolyAI/minds14) hat zum Beispiel eine Abtastrate von 8000 kHz. Um das Wav2Vec2-Modell mit diesem Datensatz verwenden zu kรถnnen, mรผssen Sie die Abtastrate auf 16 kHz erhรถhen:
```py
>>> dataset = load_dataset("PolyAI/minds14", name="en-US", split="train")
>>> dataset[0]["audio"]
{'array': array([ 0. , 0.00024414, -0.00024414, ..., -0.00024414,
0. , 0. ], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/f14948e0e84be638dd7943ac36518a4cf3324e8b7aa331c5ab11541518e9368c/en-US~JOINT_ACCOUNT/602ba55abb1e6d0fbce92065.wav',
'sampling_rate': 8000}
```
1. Verwenden Sie die Methode [~datasets.Dataset.cast_column] von ๐ค Datasets, um die Abtastrate auf 16kHz zu erhรถhen:
```py
>>> dataset = dataset.cast_column("audio", Audio(sampling_rate=16_000))
```
2. Laden Sie die Audiodatei:
```py
>>> dataset[0]["audio"]
{'array': array([ 2.3443763e-05, 2.1729663e-04, 2.2145823e-04, ...,
3.8356509e-05, -7.3497440e-06, -2.1754686e-05], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/f14948e0e84be638dd7943ac36518a4cf3324e8b7aa331c5ab11541518e9368c/en-US~JOINT_ACCOUNT/602ba55abb1e6d0fbce92065.wav',
'sampling_rate': 16000}
```
Wie Sie sehen kรถnnen, ist die Abtastrate jetzt 16kHz!
### Merkmalsextraktor
Der nรคchste Schritt ist das Laden eines Merkmalsextraktors, um die Eingabe zu normalisieren und aufzufรผllen. Beim Auffรผllen von Textdaten wird fรผr kรผrzere Sequenzen ein `0` hinzugefรผgt. Die gleiche Idee gilt fรผr Audiodaten, und der Audio-Feature-Extraktor fรผgt eine `0` - interpretiert als Stille - zu `array` hinzu.
Laden Sie den Merkmalsextraktor mit [`AutoFeatureExtractor.from_pretrained`]:
```py
>>> from transformers import AutoFeatureExtractor
>>> feature_extractor = AutoFeatureExtractor.from_pretrained("facebook/wav2vec2-base")
```
รbergeben Sie das Audio-"Array" an den Feature-Extraktor. Wir empfehlen auch, das Argument `sampling_rate` im Feature Extractor hinzuzufรผgen, um eventuell auftretende stille Fehler besser zu beheben.
```py
>>> audio_input = [dataset[0]["audio"]["array"]]
>>> feature_extractor(audio_input, sampling_rate=16000)
{'input_values': [array([ 3.8106556e-04, 2.7506407e-03, 2.8015103e-03, ...,
5.6335266e-04, 4.6588284e-06, -1.7142107e-04], dtype=float32)]}
```
### Auffรผllen und Kรผrzen
Genau wie beim Tokenizer kรถnnen Sie variable Sequenzen in einem Stapel durch Auffรผllen oder Abschneiden behandeln. Werfen Sie einen Blick auf die Sequenzlรคnge dieser beiden Audiobeispiele:
```py
>>> dataset[0]["audio"]["array"].shape
(173398,)
>>> dataset[1]["audio"]["array"].shape
(106496,)
```
Wie Sie sehen kรถnnen, hat das erste Beispiel eine lรคngere Sequenz als das zweite Beispiel. Lassen Sie uns eine Funktion erstellen, die den Datensatz vorverarbeitet. Geben Sie eine maximale Lรคnge der Probe an, und der Feature-Extraktor wird die Sequenzen entweder auffรผllen oder abschneiden, damit sie dieser Lรคnge entsprechen:
```py
>>> def preprocess_function(examples):
... audio_arrays = [x["array"] for x in examples["audio"]]
... inputs = feature_extractor(
... audio_arrays,
... sampling_rate=16000,
... padding=True,
... max_length=100000,
... truncation=True,
... )
... return inputs
```
Wenden Sie die Funktion auf die ersten paar Beispiele im Datensatz an:
```py
>>> processed_dataset = preprocess_function(dataset[:5])
```
Schauen Sie sich nun noch einmal die verarbeiteten Beispiel-Lรคngen an:
```py
>>> processed_dataset["input_values"][0].shape
(100000,)
>>> processed_dataset["input_values"][1].shape
(100000,)
```
Die Lรคnge der ersten beiden Beispiele entspricht nun der von Ihnen angegebenen Maximallรคnge.
## Bildverarbeitung
Ein Merkmalsextraktor wird auch verwendet, um Bilder fรผr Bildverarbeitungsaufgaben zu verarbeiten. Auch hier besteht das Ziel darin, das Rohbild in eine Reihe von Tensoren als Eingabe zu konvertieren.
Laden wir den [food101](https://huggingface.co/datasets/food101) Datensatz fรผr dieses Tutorial. Verwenden Sie den Parameter ๐ค Datasets `split`, um nur eine kleine Stichprobe aus dem Trainingssplit zu laden, da der Datensatz recht groร ist:
```py
>>> from datasets import load_dataset
>>> dataset = load_dataset("food101", split="train[:100]")
```
Als Nรคchstes sehen Sie sich das Bild mit dem Merkmal ๐ค Datensรคtze [Bild] (https://huggingface.co/docs/datasets/package_reference/main_classes?highlight=image#datasets.Image) an:
```py
>>> dataset[0]["image"]
```
### Merkmalsextraktor
Laden Sie den Merkmalsextraktor mit [`AutoImageProcessor.from_pretrained`]:
```py
>>> from transformers import AutoImageProcessor
>>> image_processor = AutoImageProcessor.from_pretrained("google/vit-base-patch16-224")
```
### Datenerweiterung
Bei Bildverarbeitungsaufgaben ist es รผblich, den Bildern als Teil der Vorverarbeitung eine Art von Datenerweiterung hinzuzufรผgen. Sie kรถnnen Erweiterungen mit jeder beliebigen Bibliothek hinzufรผgen, aber in diesem Tutorial werden Sie das Modul [`transforms`](https://pytorch.org/vision/stable/transforms.html) von torchvision verwenden.
1. Normalisieren Sie das Bild und verwenden Sie [`Compose`](https://pytorch.org/vision/master/generated/torchvision.transforms.Compose.html), um einige Transformationen - [`RandomResizedCrop`](https://pytorch.org/vision/main/generated/torchvision.transforms.RandomResizedCrop.html) und [`ColorJitter`](https://pytorch.org/vision/main/generated/torchvision.transforms.ColorJitter.html) - miteinander zu verknรผpfen:
```py
>>> from torchvision.transforms import Compose, Normalize, RandomResizedCrop, ColorJitter, ToTensor
>>> normalize = Normalize(mean=image_processor.image_mean, std=image_processor.image_std)
>>> _transforms = Compose(
... [RandomResizedCrop(image_processor.size["height"]), ColorJitter(brightness=0.5, hue=0.5), ToTensor(), normalize]
... )
```
2. Das Modell akzeptiert [`pixel_values`](model_doc/visionencoderdecoder#transformers.VisionEncoderDecoderModel.forward.pixel_values) als Eingabe. Dieser Wert wird vom Merkmalsextraktor erzeugt. Erstellen Sie eine Funktion, die `pixel_values` aus den Transformationen erzeugt:
```py
>>> def transforms(examples):
... examples["pixel_values"] = [_transforms(image.convert("RGB")) for image in examples["image"]]
... return examples
```
3. Dann verwenden Sie ๐ค Datasets [`set_transform`](https://huggingface.co/docs/datasets/process#format-transform), um die Transformationen im laufenden Betrieb anzuwenden:
```py
>>> dataset.set_transform(transforms)
```
4. Wenn Sie nun auf das Bild zugreifen, werden Sie feststellen, dass der Feature Extractor die Modelleingabe "pixel_values" hinzugefรผgt hat:
```py
>>> dataset[0]["image"]
{'image': <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=384x512 at 0x7F1A7B0630D0>,
'label': 6,
'pixel_values': tensor([[[ 0.0353, 0.0745, 0.1216, ..., -0.9922, -0.9922, -0.9922],
[-0.0196, 0.0667, 0.1294, ..., -0.9765, -0.9843, -0.9922],
[ 0.0196, 0.0824, 0.1137, ..., -0.9765, -0.9686, -0.8667],
...,
[ 0.0275, 0.0745, 0.0510, ..., -0.1137, -0.1216, -0.0824],
[ 0.0667, 0.0824, 0.0667, ..., -0.0588, -0.0745, -0.0980],
[ 0.0353, 0.0353, 0.0431, ..., -0.0039, -0.0039, -0.0588]],
[[ 0.2078, 0.2471, 0.2863, ..., -0.9451, -0.9373, -0.9451],
[ 0.1608, 0.2471, 0.3098, ..., -0.9373, -0.9451, -0.9373],
[ 0.2078, 0.2706, 0.3020, ..., -0.9608, -0.9373, -0.8275],
...,
[-0.0353, 0.0118, -0.0039, ..., -0.2392, -0.2471, -0.2078],
[ 0.0196, 0.0353, 0.0196, ..., -0.1843, -0.2000, -0.2235],
[-0.0118, -0.0039, -0.0039, ..., -0.0980, -0.0980, -0.1529]],
[[ 0.3961, 0.4431, 0.4980, ..., -0.9216, -0.9137, -0.9216],
[ 0.3569, 0.4510, 0.5216, ..., -0.9059, -0.9137, -0.9137],
[ 0.4118, 0.4745, 0.5216, ..., -0.9137, -0.8902, -0.7804],
...,
[-0.2314, -0.1922, -0.2078, ..., -0.4196, -0.4275, -0.3882],
[-0.1843, -0.1686, -0.2000, ..., -0.3647, -0.3804, -0.4039],
[-0.1922, -0.1922, -0.1922, ..., -0.2941, -0.2863, -0.3412]]])}
```
Hier sehen Sie, wie das Bild nach der Vorverarbeitung aussieht. Wie von den angewandten Transformationen zu erwarten, wurde das Bild willkรผrlich beschnitten und seine Farbeigenschaften sind anders.
```py
>>> import numpy as np
>>> import matplotlib.pyplot as plt
>>> img = dataset[0]["pixel_values"]
>>> plt.imshow(img.permute(1, 2, 0))
```
## Multimodal
Fรผr multimodale Aufgaben werden Sie eine Kombination aus allem, was Sie bisher gelernt haben, verwenden und Ihre Fรคhigkeiten auf eine Aufgabe der automatischen Spracherkennung (ASR) anwenden. Dies bedeutet, dass Sie einen:
* Feature Extractor zur Vorverarbeitung der Audiodaten.
* Tokenizer, um den Text zu verarbeiten.
Kehren wir zum [LJ Speech](https://huggingface.co/datasets/lj_speech) Datensatz zurรผck:
```py
>>> from datasets import load_dataset
>>> lj_speech = load_dataset("lj_speech", split="train")
```
Da Sie hauptsรคchlich an den Spalten "Audio" und "Text" interessiert sind, entfernen Sie die anderen Spalten:
```py
>>> lj_speech = lj_speech.map(remove_columns=["file", "id", "normalized_text"])
```
Schauen Sie sich nun die Spalten "Audio" und "Text" an:
```py
>>> lj_speech[0]["audio"]
{'array': array([-7.3242188e-04, -7.6293945e-04, -6.4086914e-04, ...,
7.3242188e-04, 2.1362305e-04, 6.1035156e-05], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/917ece08c95cf0c4115e45294e3cd0dee724a1165b7fc11798369308a465bd26/LJSpeech-1.1/wavs/LJ001-0001.wav',
'sampling_rate': 22050}
>>> lj_speech[0]["text"]
'Printing, in the only sense with which we are at present concerned, differs from most if not from all the arts and crafts represented in the Exhibition'
```
Erinnern Sie sich an den frรผheren Abschnitt รผber die Verarbeitung von Audiodaten: Sie sollten immer die Abtastrate Ihrer Audiodaten [resample](preprocessing#audio), damit sie mit der Abtastrate des Datensatzes รผbereinstimmt, der fรผr das Vortraining eines Modells verwendet wird:
```py
>>> lj_speech = lj_speech.cast_column("audio", Audio(sampling_rate=16_000))
```
### Prozessor
Ein Processor kombiniert einen Feature-Extraktor und einen Tokenizer. Laden Sie einen Processor mit [`AutoProcessor.from_pretrained]:
```py
>>> from transformers import AutoProcessor
>>> processor = AutoProcessor.from_pretrained("facebook/wav2vec2-base-960h")
```
1. Erstellen Sie eine Funktion, die die Audiodaten zu `input_values` verarbeitet und den Text zu `labels` tokenisiert. Dies sind Ihre Eingaben fรผr das Modell:
```py
>>> def prepare_dataset(example):
... audio = example["audio"]
... example.update(processor(audio=audio["array"], text=example["text"], sampling_rate=16000))
... return example
```
2. Wenden Sie die Funktion "prepare_dataset" auf ein Beispiel an:
```py
>>> prepare_dataset(lj_speech[0])
```
Beachten Sie, dass der Processor `input_values` und `labels` hinzugefรผgt hat. Auch die Abtastrate wurde korrekt auf 16kHz heruntergerechnet.
Toll, Sie sollten jetzt in der Lage sein, Daten fรผr jede Modalitรคt vorzuverarbeiten und sogar verschiedene Modalitรคten zu kombinieren! Im nรคchsten Kurs lernen Sie, wie Sie ein Modell mit Ihren neu aufbereiteten Daten feinabstimmen kรถnnen.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/model_sharing.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Ein Modell teilen
Die letzten beiden Tutorials haben gezeigt, wie man ein Modell mit PyTorch, Keras und ๐ค Accelerate fรผr verteilte Setups feinabstimmen kann. Der nรคchste Schritt besteht darin, Ihr Modell mit der Community zu teilen! Bei Hugging Face glauben wir an den offenen Austausch von Wissen und Ressourcen, um kรผnstliche Intelligenz fรผr alle zu demokratisieren. Wir ermutigen Sie, Ihr Modell mit der Community zu teilen, um anderen zu helfen, Zeit und Ressourcen zu sparen.
In diesem Tutorial lernen Sie zwei Methoden kennen, wie Sie ein trainiertes oder verfeinertes Modell auf dem [Model Hub](https://huggingface.co/models) teilen kรถnnen:
- Programmgesteuertes รbertragen Ihrer Dateien auf den Hub.
- Ziehen Sie Ihre Dateien per Drag-and-Drop รผber die Weboberflรคche in den Hub.
<iframe width="560" height="315" src="https://www.youtube.com/embed/XvSGPZFEjDY" title="YouTube video player"
frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope;
picture-in-picture" allowfullscreen></iframe>
<Tip>
Um ein Modell mit der รffentlichkeit zu teilen, benรถtigen Sie ein Konto auf [huggingface.co](https://huggingface.co/join). Sie kรถnnen auch einer bestehenden Organisation beitreten oder eine neue Organisation grรผnden.
</Tip>
## Repository-Funktionen
Jedes Repository im Model Hub verhรคlt sich wie ein typisches GitHub-Repository. Unsere Repositorys bieten Versionierung, Commit-Historie und die Mรถglichkeit, Unterschiede zu visualisieren.
Die integrierte Versionierung des Model Hub basiert auf Git und [git-lfs](https://git-lfs.github.com/). Mit anderen Worten: Sie kรถnnen ein Modell als ein Repository behandeln, was eine bessere Zugriffskontrolle und Skalierbarkeit ermรถglicht. Die Versionskontrolle ermรถglicht *Revisionen*, eine Methode zum Anheften einer bestimmten Version eines Modells mit einem Commit-Hash, Tag oder Branch.
Folglich kรถnnen Sie eine bestimmte Modellversion mit dem Parameter "Revision" laden:
```py
>>> model = AutoModel.from_pretrained(
... "julien-c/EsperBERTo-small", revision="v2.0.1" # tag name, or branch name, or commit hash
... )
```
Dateien lassen sich auch in einem Repository leicht bearbeiten, und Sie kรถnnen die Commit-Historie sowie die Unterschiede einsehen:
## Einrichtung
Bevor Sie ein Modell fรผr den Hub freigeben, benรถtigen Sie Ihre Hugging Face-Anmeldedaten. Wenn Sie Zugang zu einem Terminal haben, fรผhren Sie den folgenden Befehl in der virtuellen Umgebung aus, in der ๐ค Transformers installiert ist. Dadurch werden Ihre Zugangsdaten in Ihrem Hugging Face-Cache-Ordner (standardmรครig `~/.cache/`) gespeichert:
```bash
huggingface-cli login
```
Wenn Sie ein Notebook wie Jupyter oder Colaboratory verwenden, stellen Sie sicher, dass Sie die [`huggingface_hub`](https://huggingface.co/docs/hub/adding-a-library) Bibliothek installiert haben. Diese Bibliothek ermรถglicht Ihnen die programmatische Interaktion mit dem Hub.
```bash
pip install huggingface_hub
```
Verwenden Sie dann `notebook_login`, um sich beim Hub anzumelden, und folgen Sie dem Link [hier](https://huggingface.co/settings/token), um ein Token fรผr die Anmeldung zu generieren:
```py
>>> from huggingface_hub import notebook_login
>>> notebook_login()
```
## Ein Modell fรผr alle Frameworks konvertieren
Um sicherzustellen, dass Ihr Modell von jemandem verwendet werden kann, der mit einem anderen Framework arbeitet, empfehlen wir Ihnen, Ihr Modell sowohl mit PyTorch- als auch mit TensorFlow-Checkpoints zu konvertieren und hochzuladen. Wรคhrend Benutzer immer noch in der Lage sind, Ihr Modell von einem anderen Framework zu laden, wenn Sie diesen Schritt รผberspringen, wird es langsamer sein, weil ๐ค Transformers den Checkpoint on-the-fly konvertieren mรผssen.
Die Konvertierung eines Checkpoints fรผr ein anderes Framework ist einfach. Stellen Sie sicher, dass Sie PyTorch und TensorFlow installiert haben (siehe [hier](installation) fรผr Installationsanweisungen), und finden Sie dann das spezifische Modell fรผr Ihre Aufgabe in dem anderen Framework.
<frameworkcontent>
<pt>
Geben Sie `from_tf=True` an, um einen Prรผfpunkt von TensorFlow nach PyTorch zu konvertieren:
```py
>>> pt_model = DistilBertForSequenceClassification.from_pretrained("path/to/awesome-name-you-picked", from_tf=True)
>>> pt_model.save_pretrained("path/to/awesome-name-you-picked")
```
</pt>
<tf>
Geben Sie `from_pt=True` an, um einen Prรผfpunkt von PyTorch nach TensorFlow zu konvertieren:
```py
>>> tf_model = TFDistilBertForSequenceClassification.from_pretrained("path/to/awesome-name-you-picked", from_pt=True)
```
Dann kรถnnen Sie Ihr neues TensorFlow-Modell mit seinem neuen Checkpoint speichern:
```py
>>> tf_model.save_pretrained("path/to/awesome-name-you-picked")
```
</tf>
<jax>
Wenn ein Modell in Flax verfรผgbar ist, kรถnnen Sie auch einen Kontrollpunkt von PyTorch nach Flax konvertieren:
```py
>>> flax_model = FlaxDistilBertForSequenceClassification.from_pretrained(
... "path/to/awesome-name-you-picked", from_pt=True
... )
```
</jax>
</frameworkcontent>
## Ein Modell wรคhrend des Trainings hochladen
<frameworkcontent>
<pt>
<Youtube id="Z1-XMy-GNLQ"/>
Die Weitergabe eines Modells an den Hub ist so einfach wie das Hinzufรผgen eines zusรคtzlichen Parameters oder Rรผckrufs. Erinnern Sie sich an das [Feinabstimmungs-Tutorial](training), in der Klasse [`TrainingArguments`] geben Sie Hyperparameter und zusรคtzliche Trainingsoptionen an. Eine dieser Trainingsoptionen beinhaltet die Mรถglichkeit, ein Modell direkt an den Hub zu pushen. Setzen Sie `push_to_hub=True` in Ihrer [`TrainingArguments`]:
```py
>>> training_args = TrainingArguments(output_dir="my-awesome-model", push_to_hub=True)
```
รbergeben Sie Ihre Trainingsargumente wie gewohnt an [`Trainer`]:
```py
>>> trainer = Trainer(
... model=model,
... args=training_args,
... train_dataset=small_train_dataset,
... eval_dataset=small_eval_dataset,
... compute_metrics=compute_metrics,
... )
```
Nach der Feinabstimmung Ihres Modells rufen Sie [`~transformers.Trainer.push_to_hub`] auf [`Trainer`] auf, um das trainierte Modell an den Hub zu รผbertragen. Transformers fรผgt sogar automatisch Trainings-Hyperparameter, Trainingsergebnisse und Framework-Versionen zu Ihrer Modellkarte hinzu!
```py
>>> trainer.push_to_hub()
```
</pt>
<tf>
Geben Sie ein Modell mit [`PushToHubCallback`] an den Hub weiter. In der [`PushToHubCallback`] Funktion, fรผgen Sie hinzu:
- Ein Ausgabeverzeichnis fรผr Ihr Modell.
- Einen Tokenizer.
- Die `hub_model_id`, die Ihr Hub-Benutzername und Modellname ist.
```py
>>> from transformers import PushToHubCallback
>>> push_to_hub_callback = PushToHubCallback(
... output_dir="./your_model_save_path", tokenizer=tokenizer, hub_model_id="your-username/my-awesome-model"
... )
```
Fรผgen Sie den Callback zu [`fit`](https://keras.io/api/models/model_training_apis/) hinzu, und ๐ค Transformers wird das trainierte Modell an den Hub weiterleiten:
```py
>>> model.fit(tf_train_dataset, validation_data=tf_validation_dataset, epochs=3, callbacks=push_to_hub_callback)
```
</tf>
</frameworkcontent>
## Verwenden Sie die Funktion `push_to_hub`.
Sie kรถnnen `push_to_hub` auch direkt fรผr Ihr Modell aufrufen, um es in den Hub hochzuladen.
Geben Sie den Namen Ihres Modells in "push_to_hub" an:
```py
>>> pt_model.push_to_hub("my-awesome-model")
```
Dadurch wird ein Repository unter Ihrem Benutzernamen mit dem Modellnamen `my-awesome-model` erstellt. Benutzer kรถnnen nun Ihr Modell mit der Funktion `from_pretrained` laden:
```py
>>> from transformers import AutoModel
>>> model = AutoModel.from_pretrained("your_username/my-awesome-model")
```
Wenn Sie zu einer Organisation gehรถren und Ihr Modell stattdessen unter dem Namen der Organisation pushen wollen, fรผgen Sie diesen einfach zur `repo_id` hinzu:
```py
>>> pt_model.push_to_hub("my-awesome-org/my-awesome-model")
```
Die Funktion "push_to_hub" kann auch verwendet werden, um andere Dateien zu einem Modell-Repository hinzuzufรผgen. Zum Beispiel kann man einen Tokenizer zu einem Modell-Repository hinzufรผgen:
```py
>>> tokenizer.push_to_hub("my-awesome-model")
```
Oder vielleicht mรถchten Sie die TensorFlow-Version Ihres fein abgestimmten PyTorch-Modells hinzufรผgen:
```py
>>> tf_model.push_to_hub("my-awesome-model")
```
Wenn Sie nun zu Ihrem Hugging Face-Profil navigieren, sollten Sie Ihr neu erstelltes Modell-Repository sehen. Wenn Sie auf die Registerkarte **Dateien** klicken, werden alle Dateien angezeigt, die Sie in das Repository hochgeladen haben.
Weitere Einzelheiten zum Erstellen und Hochladen von Dateien in ein Repository finden Sie in der Hub-Dokumentation [hier](https://huggingface.co/docs/hub/how-to-upstream).
## Hochladen mit der Weboberflรคche
Benutzer, die einen no-code Ansatz bevorzugen, kรถnnen ein Modell รผber das Webinterface des Hubs hochladen. Besuchen Sie [huggingface.co/new](https://huggingface.co/new) um ein neues Repository zu erstellen:
Fรผgen Sie von hier aus einige Informationen รผber Ihr Modell hinzu:
- Wรคhlen Sie den **Besitzer** des Repositorys. Dies kรถnnen Sie selbst oder eine der Organisationen sein, denen Sie angehรถren.
- Wรคhlen Sie einen Namen fรผr Ihr Modell, der auch der Name des Repositorys sein wird.
- Wรคhlen Sie, ob Ihr Modell รถffentlich oder privat ist.
- Geben Sie die Lizenzverwendung fรผr Ihr Modell an.
Klicken Sie nun auf die Registerkarte **Dateien** und klicken Sie auf die Schaltflรคche **Datei hinzufรผgen**, um eine neue Datei in Ihr Repository hochzuladen. Ziehen Sie dann eine Datei per Drag-and-Drop hoch und fรผgen Sie eine รbergabemeldung hinzu.
## Hinzufรผgen einer Modellkarte
Um sicherzustellen, dass die Benutzer die Fรคhigkeiten, Grenzen, mรถglichen Verzerrungen und ethischen Aspekte Ihres Modells verstehen, fรผgen Sie bitte eine Modellkarte zu Ihrem Repository hinzu. Die Modellkarte wird in der Datei `README.md` definiert. Sie kรถnnen eine Modellkarte hinzufรผgen, indem Sie:
* Manuelles Erstellen und Hochladen einer "README.md"-Datei.
* Klicken Sie auf die Schaltflรคche **Modellkarte bearbeiten** in Ihrem Modell-Repository.
Werfen Sie einen Blick auf die DistilBert [model card](https://huggingface.co/distilbert-base-uncased) als gutes Beispiel fรผr die Art von Informationen, die eine Modellkarte enthalten sollte. Weitere Details รผber andere Optionen, die Sie in der Datei "README.md" einstellen kรถnnen, wie z.B. den Kohlenstoff-Fuรabdruck eines Modells oder Beispiele fรผr Widgets, finden Sie in der Dokumentation [hier](https://huggingface.co/docs/hub/models-cards).
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/installation.md
|
<!---
Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Installation
Installieren Sie ๐ค Transformers fรผr die Deep-Learning-Bibliothek, mit der Sie arbeiten, richten Sie Ihren Cache ein und konfigurieren Sie ๐ค Transformers optional fรผr den Offline-Betrieb.
๐ค Transformers wurde unter Python 3.6+, PyTorch 1.1.0+, TensorFlow 2.0+, und Flax getestet. Folgen Sie den Installationsanweisungen unten fรผr die von Ihnen verwendete Deep-Learning-Bibliothek:
* [PyTorch](https://pytorch.org/get-started/locally/) installation instructions.
* [TensorFlow 2.0](https://www.tensorflow.org/install/pip) installation instructions.
* [Flax](https://flax.readthedocs.io/en/latest/) installation instructions.
## Installation mit pip
Sie sollten ๐ค Transformers in einer [virtuellen Umgebung](https://docs.python.org/3/library/venv.html) installieren. Wenn Sie mit virtuellen Python-Umgebungen nicht vertraut sind, werfen Sie einen Blick auf diese [Anleitung](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/). Eine virtuelle Umgebung macht es einfacher, verschiedene Projekte zu verwalten und Kompatibilitรคtsprobleme zwischen Abhรคngigkeiten zu vermeiden.
Beginnen wir mit der Erstellung einer virtuellen Umgebung in Ihrem Projektverzeichnis:
```bash
python -m venv .env
```
Aktivieren wir die virtuelle Umgebung. Unter Linux und MacOs:
```bash
source .env/bin/activate
```
Aktivieren wir die virtuelle Umgebung unter Windows
```bash
.env/Scripts/activate
```
Jetzt kรถnnen wir die ๐ค Transformers mit dem folgenden Befehl installieren:
```bash
pip install transformers
```
Bei reiner CPU-Unterstรผtzung kรถnnen wir ๐ค Transformers und eine Deep-Learning-Bibliothek bequem in einer Zeile installieren. Installieren wir zum Beispiel ๐ค Transformers und PyTorch mit:
```bash
pip install transformers[torch]
```
๐ค Transformers und TensorFlow 2.0:
```bash
pip install transformers[tf-cpu]
```
๐ค Transformers und Flax:
```bash
pip install transformers[flax]
```
รberprรผfen wir abschlieรend, ob ๐ค Transformers ordnungsgemรคร installiert wurde, indem wir den folgenden Befehl ausfรผhren. Es wird ein vortrainiertes Modell heruntergeladen:
```bash
python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('we love you'))"
```
Dann wird die Kategorie und die Wahrscheinlichkeit ausgegeben:
```bash
[{'label': 'POSITIVE', 'score': 0.9998704791069031}]
```
## Installation aus dem Code
Installieren wir ๐ค Transformers aus dem Quellcode mit dem folgenden Befehl:
```bash
pip install git+https://github.com/huggingface/transformers
```
Dieser Befehl installiert die aktuelle `main` Version und nicht die neueste `stable` Version. Die `main`-Version ist nรผtzlich, um mit den neuesten Entwicklungen Schritt zu halten. Zum Beispiel, wenn ein Fehler seit der letzten offiziellen Version behoben wurde, aber eine neue Version noch nicht verรถffentlicht wurde. Das bedeutet jedoch, dass die "Hauptversion" nicht immer stabil ist. Wir bemรผhen uns, die Hauptversion einsatzbereit zu halten, und die meisten Probleme werden normalerweise innerhalb weniger Stunden oder eines Tages behoben. Wenn Sie auf ein Problem stoรen, รถffnen Sie bitte ein [Issue] (https://github.com/huggingface/transformers/issues), damit wir es noch schneller beheben kรถnnen!
รberprรผfen wir, ob ๐ค Transformers richtig installiert wurde, indem Sie den folgenden Befehl ausfรผhren:
```bash
python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('I love you'))"
```
## Editierbare Installation
Sie benรถtigen eine bearbeitbare Installation, wenn Sie:
* die "Haupt"-Version des Quellcodes verwenden mรถchten.
* Zu ๐ค Transformers beitragen und รnderungen am Code testen wollen.
Klonen Sie das Repository und installieren ๐ค Transformers mit den folgenden Befehlen:
```bash
git clone https://github.com/huggingface/transformers.git
cd transformers
pip install -e .
```
Diese Befehle verknรผpfen den Ordner, in den Sie das Repository geklont haben, mit den Pfaden Ihrer Python-Bibliotheken. Python wird nun in dem Ordner suchen, in den Sie geklont haben, zusรคtzlich zu den normalen Bibliothekspfaden. Wenn zum Beispiel Ihre Python-Pakete normalerweise in `~/anaconda3/envs/main/lib/python3.7/site-packages/` installiert sind, wird Python auch den Ordner durchsuchen, in den Sie geklont haben: `~/transformers/`.
<Tip warning={true}>
Sie mรผssen den Ordner `transformers` behalten, wenn Sie die Bibliothek weiter verwenden wollen.
</Tip>
Jetzt kรถnnen Sie Ihren Klon mit dem folgenden Befehl ganz einfach auf die neueste Version von ๐ค Transformers aktualisieren:
```bash
cd ~/transformers/
git pull
```
Ihre Python-Umgebung wird beim nรคchsten Ausfรผhren die `main`-Version von ๐ค Transformers finden.
## Installation mit conda
Installation von dem conda Kanal `huggingface`:
```bash
conda install -c huggingface transformers
```
## Cache Einrichtung
Vorgefertigte Modelle werden heruntergeladen und lokal zwischengespeichert unter: `~/.cache/huggingface/hub`. Dies ist das Standardverzeichnis, das durch die Shell-Umgebungsvariable "TRANSFORMERS_CACHE" vorgegeben ist. Unter Windows wird das Standardverzeichnis durch `C:\Benutzer\Benutzername\.cache\huggingface\hub` angegeben. Sie kรถnnen die unten aufgefรผhrten Shell-Umgebungsvariablen - in der Reihenfolge ihrer Prioritรคt - รคndern, um ein anderes Cache-Verzeichnis anzugeben:
1. Shell-Umgebungsvariable (Standard): `HUGGINGFACE_HUB_CACHE` oder `TRANSFORMERS_CACHE`.
2. Shell-Umgebungsvariable: `HF_HOME`.
3. Shell-Umgebungsvariable: `XDG_CACHE_HOME` + `/huggingface`.
<Tip>
Transformers verwendet die Shell-Umgebungsvariablen `PYTORCH_TRANSFORMERS_CACHE` oder `PYTORCH_PRETRAINED_BERT_CACHE`, wenn Sie von einer frรผheren Iteration dieser Bibliothek kommen und diese Umgebungsvariablen gesetzt haben, sofern Sie nicht die Shell-Umgebungsvariable `TRANSFORMERS_CACHE` angeben.
</Tip>
## Offline Modus
Transformers ist in der Lage, in einer Firewall- oder Offline-Umgebung zu laufen, indem es nur lokale Dateien verwendet. Setzen Sie die Umgebungsvariable `TRANSFORMERS_OFFLINE=1`, um dieses Verhalten zu aktivieren.
<Tip>
Fรผgen sie [๐ค Datasets](https://huggingface.co/docs/datasets/) zu Ihrem Offline-Trainingsworkflow hinzufรผgen, indem Sie die Umgebungsvariable `HF_DATASETS_OFFLINE=1` setzen.
</Tip>
So wรผrden Sie beispielsweise ein Programm in einem normalen Netzwerk mit einer Firewall fรผr externe Instanzen mit dem folgenden Befehl ausfรผhren:
```bash
python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
```
Fรผhren Sie das gleiche Programm in einer Offline-Instanz mit aus:
```bash
HF_DATASETS_OFFLINE=1 TRANSFORMERS_OFFLINE=1 \
python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
```
Das Skript sollte nun laufen, ohne sich aufzuhรคngen oder eine Zeitรผberschreitung abzuwarten, da es weiร, dass es nur nach lokalen Dateien suchen soll.
### Abrufen von Modellen und Tokenizern zur Offline-Verwendung
Eine andere Mรถglichkeit, ๐ค Transformers offline zu verwenden, besteht darin, die Dateien im Voraus herunterzuladen und dann auf ihren lokalen Pfad zu verweisen, wenn Sie sie offline verwenden mรผssen. Es gibt drei Mรถglichkeiten, dies zu tun:
* Laden Sie eine Datei รผber die Benutzeroberflรคche des [Model Hub](https://huggingface.co/models) herunter, indem Sie auf das โ-Symbol klicken.
* Verwenden Sie den [PreTrainedModel.from_pretrained] und [PreTrainedModel.save_pretrained] Workflow:
1. Laden Sie Ihre Dateien im Voraus mit [`PreTrainedModel.from_pretrained`] herunter:
```py
>>> from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
>>> tokenizer = AutoTokenizer.from_pretrained("bigscience/T0_3B")
>>> model = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0_3B")
```
2. Speichern Sie Ihre Dateien in einem bestimmten Verzeichnis mit [`PreTrainedModel.save_pretrained`]:
```py
>>> tokenizer.save_pretrained("./your/path/bigscience_t0")
>>> model.save_pretrained("./your/path/bigscience_t0")
```
3. Wenn Sie nun offline sind, laden Sie Ihre Dateien mit [`PreTrainedModel.from_pretrained`] aus dem bestimmten Verzeichnis:
```py
>>> tokenizer = AutoTokenizer.from_pretrained("./your/path/bigscience_t0")
>>> model = AutoModel.from_pretrained("./your/path/bigscience_t0")
```
* Programmatisches Herunterladen von Dateien mit der [huggingface_hub](https://github.com/huggingface/huggingface_hub/tree/main/src/huggingface_hub) Bibliothek:
1. Installieren Sie die "huggingface_hub"-Bibliothek in Ihrer virtuellen Umgebung:
```bash
python -m pip install huggingface_hub
```
2. Verwenden Sie die Funktion [`hf_hub_download`](https://huggingface.co/docs/hub/adding-a-library#download-files-from-the-hub), um eine Datei in einen bestimmten Pfad herunterzuladen. Der folgende Befehl lรคdt zum Beispiel die Datei "config.json" aus dem Modell [T0](https://huggingface.co/bigscience/T0_3B) in den gewรผnschten Pfad herunter:
```py
>>> from huggingface_hub import hf_hub_download
>>> hf_hub_download(repo_id="bigscience/T0_3B", filename="config.json", cache_dir="./your/path/bigscience_t0")
```
Sobald Ihre Datei heruntergeladen und lokal zwischengespeichert ist, geben Sie den lokalen Pfad an, um sie zu laden und zu verwenden:
```py
>>> from transformers import AutoConfig
>>> config = AutoConfig.from_pretrained("./your/path/bigscience_t0/config.json")
```
<Tip>
Weitere Informationen zum Herunterladen von Dateien, die auf dem Hub gespeichert sind, finden Sie im Abschnitt [Wie man Dateien vom Hub herunterlรคdt] (https://huggingface.co/docs/hub/how-to-downstream).
</Tip>
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/pipeline_tutorial.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Pipelines fรผr Inferenzen
Die [`pipeline`] macht es einfach, jedes beliebige Modell aus dem [Hub](https://huggingface.co/models) fรผr die Inferenz auf jede Sprache, Computer Vision, Sprache und multimodale Aufgaben zu verwenden. Selbst wenn Sie keine Erfahrung mit einer bestimmten Modalitรคt haben oder nicht mit dem zugrundeliegenden Code hinter den Modellen vertraut sind, kรถnnen Sie sie mit der [`pipeline`] fรผr Inferenzen verwenden! In diesem Beispiel lernen Sie, wie:
* Eine [`pipeline`] fรผr Inferenz zu verwenden.
* Einen bestimmten Tokenizer oder ein bestimmtes Modell zu verwenden.
* Eine [`pipeline`] fรผr Audio-, Vision- und multimodale Aufgaben zu verwenden.
<Tip>
Eine vollstรคndige Liste der unterstรผtzten Aufgaben und verfรผgbaren Parameter finden Sie in der [`pipeline`]-Dokumentation.
</Tip>
## Verwendung von Pipelines
Obwohl jede Aufgabe eine zugehรถrige [`pipeline`] hat, ist es einfacher, die allgemeine [`pipeline`]-Abstraktion zu verwenden, die alle aufgabenspezifischen Pipelines enthรคlt. Die [`pipeline`] lรคdt automatisch ein Standardmodell und eine Vorverarbeitungsklasse, die fรผr Ihre Aufgabe inferenzfรคhig ist.
1. Beginnen Sie mit der Erstellung einer [`pipeline`] und geben Sie eine Inferenzaufgabe an:
```py
>>> from transformers import pipeline
>>> generator = pipeline(task="text-generation")
```
2. รbergeben Sie Ihren Eingabetext an die [`pipeline`]:
```py
>>> generator(
... "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone"
... ) # doctest: +SKIP
[{'generated_text': 'Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone, Seven for the Iron-priests at the door to the east, and thirteen for the Lord Kings at the end of the mountain'}]
```
Wenn Sie mehr als eine Eingabe haben, รผbergeben Sie die Eingabe als Liste:
```py
>>> generator(
... [
... "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone",
... "Nine for Mortal Men, doomed to die, One for the Dark Lord on his dark throne",
... ]
... ) # doctest: +SKIP
```
Alle zusรคtzlichen Parameter fรผr Ihre Aufgabe kรถnnen auch in die [`pipeline`] aufgenommen werden. Die Aufgabe `Text-Generierung` hat eine [`~generation.GenerationMixin.generate`]-Methode mit mehreren Parametern zur Steuerung der Ausgabe. Wenn Sie zum Beispiel mehr als eine Ausgabe erzeugen wollen, setzen Sie den Parameter `num_return_sequences`:
```py
>>> generator(
... "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone",
... num_return_sequences=2,
... ) # doctest: +SKIP
```
### Wรคhlen Sie ein Modell und einen Tokenizer
Die [`pipeline`] akzeptiert jedes Modell aus dem [Hub] (https://huggingface.co/models). Auf dem Hub gibt es Tags, mit denen Sie nach einem Modell filtern kรถnnen, das Sie fรผr Ihre Aufgabe verwenden mรถchten. Sobald Sie ein passendes Modell ausgewรคhlt haben, laden Sie es mit der entsprechenden `AutoModelFor` und [`AutoTokenizer`] Klasse. Laden Sie zum Beispiel die Klasse [`AutoModelForCausalLM`] fรผr eine kausale Sprachmodellierungsaufgabe:
```py
>>> from transformers import AutoTokenizer, AutoModelForCausalLM
>>> tokenizer = AutoTokenizer.from_pretrained("distilgpt2")
>>> model = AutoModelForCausalLM.from_pretrained("distilgpt2")
```
Erstellen Sie eine [`pipeline`] fรผr Ihre Aufgabe, und geben Sie das Modell und den Tokenizer an, die Sie geladen haben:
```py
>>> from transformers import pipeline
>>> generator = pipeline(task="text-generation", model=model, tokenizer=tokenizer)
```
รbergeben Sie Ihren Eingabetext an die [`pipeline`] , um einen Text zu erzeugen:
```py
>>> generator(
... "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone"
... ) # doctest: +SKIP
[{'generated_text': 'Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone, Seven for the Dragon-lords (for them to rule in a world ruled by their rulers, and all who live within the realm'}]
```
## Audio-Pipeline
Die [`pipeline`] unterstรผtzt auch Audioaufgaben wie Audioklassifizierung und automatische Spracherkennung.
Lassen Sie uns zum Beispiel die Emotion in diesem Audioclip klassifizieren:
```py
>>> from datasets import load_dataset
>>> import torch
>>> torch.manual_seed(42) # doctest: +IGNORE_RESULT
>>> ds = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
>>> audio_file = ds[0]["audio"]["path"]
```
Finden Sie ein [Audioklassifikation](https://huggingface.co/models?pipeline_tag=audio-classification) Modell auf dem Model Hub fรผr Emotionserkennung und laden Sie es in die [`pipeline`]:
```py
>>> from transformers import pipeline
>>> audio_classifier = pipeline(
... task="audio-classification", model="ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition"
... )
```
รbergeben Sie die Audiodatei an die [`pipeline`]:
```py
>>> preds = audio_classifier(audio_file)
>>> preds = [{"score": round(pred["score"], 4), "label": pred["label"]} for pred in preds]
>>> preds
[{'score': 0.1315, 'label': 'calm'}, {'score': 0.1307, 'label': 'neutral'}, {'score': 0.1274, 'label': 'sad'}, {'score': 0.1261, 'label': 'fearful'}, {'score': 0.1242, 'label': 'happy'}]
```
## Bildverarbeitungs-Pipeline
Die Verwendung einer [`pipeline`] fรผr Bildverarbeitungsaufgaben ist praktisch identisch.
Geben Sie Ihre Aufgabe an und รผbergeben Sie Ihr Bild an den Klassifikator. Das Bild kann ein Link oder ein lokaler Pfad zu dem Bild sein. Zum Beispiel: Welche Katzenart ist unten abgebildet?
```py
>>> from transformers import pipeline
>>> vision_classifier = pipeline(task="image-classification")
>>> preds = vision_classifier(
... images="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/pipeline-cat-chonk.jpeg"
... )
>>> preds = [{"score": round(pred["score"], 4), "label": pred["label"]} for pred in preds]
>>> preds
[{'score': 0.4335, 'label': 'lynx, catamount'}, {'score': 0.0348, 'label': 'cougar, puma, catamount, mountain lion, painter, panther, Felis concolor'}, {'score': 0.0324, 'label': 'snow leopard, ounce, Panthera uncia'}, {'score': 0.0239, 'label': 'Egyptian cat'}, {'score': 0.0229, 'label': 'tiger cat'}]
```
## Multimodale Pipeline
Die [`pipeline`] unterstรผtzt mehr als eine Modalitรคt. Eine Aufgabe zur Beantwortung visueller Fragen (VQA) kombiniert zum Beispiel Text und Bild. Verwenden Sie einen beliebigen Bildlink und eine Frage, die Sie zu dem Bild stellen mรถchten. Das Bild kann eine URL oder ein lokaler Pfad zu dem Bild sein.
Wenn Sie zum Beispiel das gleiche Bild wie in der obigen Vision-Pipeline verwenden:
```py
>>> image = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/pipeline-cat-chonk.jpeg"
>>> question = "Where is the cat?"
```
Erstellen Sie eine Pipeline fรผr "vqa" und รผbergeben Sie ihr das Bild und die Frage:
```py
>>> from transformers import pipeline
>>> vqa = pipeline(task="vqa")
>>> preds = vqa(image=image, question=question)
>>> preds = [{"score": round(pred["score"], 4), "answer": pred["answer"]} for pred in preds]
>>> preds
[{'score': 0.9112, 'answer': 'snow'}, {'score': 0.8796, 'answer': 'in snow'}, {'score': 0.6717, 'answer': 'outside'}, {'score': 0.0291, 'answer': 'on ground'}, {'score': 0.027, 'answer': 'ground'}]
```
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/add_tensorflow_model.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Wie konvertiert man ein ๐ค Transformers-Modell in TensorFlow?
Die Tatsache, dass mehrere Frameworks fรผr die Verwendung mit ๐ค Transformers zur Verfรผgung stehen, gibt Ihnen die Flexibilitรคt, deren Stรคrken beim Entwurf Ihrer Anwendung auszuspielen.
Ihre Anwendung zu entwerfen, aber das bedeutet auch, dass die Kompatibilitรคt fรผr jedes Modell einzeln hinzugefรผgt werden muss. Die gute Nachricht ist, dass
das Hinzufรผgen von TensorFlow-Kompatibilitรคt zu einem bestehenden Modell einfacher ist als [das Hinzufรผgen eines neuen Modells von Grund auf](add_new_model)!
Ob Sie ein tieferes Verstรคndnis fรผr groรe TensorFlow-Modelle haben mรถchten, einen wichtigen Open-Source-Beitrag leisten oder
TensorFlow fรผr das Modell Ihrer Wahl aktivieren wollen, dieser Leitfaden ist fรผr Sie.
Dieser Leitfaden befรคhigt Sie, ein Mitglied unserer Gemeinschaft, TensorFlow-Modellgewichte und/oder
Architekturen beizusteuern, die in ๐ค Transformers verwendet werden sollen, und zwar mit minimaler Betreuung durch das Hugging Face Team. Das Schreiben eines neuen Modells
ist keine Kleinigkeit, aber ich hoffe, dass dieser Leitfaden dazu beitrรคgt, dass es weniger eine Achterbahnfahrt ๐ข und mehr ein Spaziergang im Park ๐ถ ist.
Die Nutzung unserer kollektiven Erfahrungen ist absolut entscheidend, um diesen Prozess immer einfacher zu machen, und deshalb mรถchten wir
ermutigen Sie daher, Verbesserungsvorschlรคge fรผr diesen Leitfaden zu machen!
Bevor Sie tiefer eintauchen, empfehlen wir Ihnen, die folgenden Ressourcen zu lesen, wenn Sie neu in ๐ค Transformers sind:
- [Allgemeiner รberblick รผber ๐ค Transformers](add_new_model#general-overview-of-transformers)
- [Die TensorFlow-Philosophie von Hugging Face](https://huggingface.co/blog/tensorflow-philosophy)
Im Rest dieses Leitfadens werden Sie lernen, was nรถtig ist, um eine neue TensorFlow Modellarchitektur hinzuzufรผgen, die
Verfahren zur Konvertierung von PyTorch in TensorFlow-Modellgewichte und wie Sie Unstimmigkeiten zwischen ML
Frameworks. Legen Sie los!
<Tip>
Sind Sie unsicher, ob das Modell, das Sie verwenden mรถchten, bereits eine entsprechende TensorFlow-Architektur hat?
รberprรผfen Sie das Feld `model_type` in der `config.json` des Modells Ihrer Wahl
([Beispiel](https://huggingface.co/bert-base-uncased/blob/main/config.json#L14)). Wenn der entsprechende Modellordner in
๐ค Transformers eine Datei hat, deren Name mit "modeling_tf" beginnt, bedeutet dies, dass es eine entsprechende TensorFlow
Architektur hat ([Beispiel](https://github.com/huggingface/transformers/tree/main/src/transformers/models/bert)).
</Tip>
## Schritt-fรผr-Schritt-Anleitung zum Hinzufรผgen von TensorFlow-Modellarchitektur-Code
Es gibt viele Mรถglichkeiten, eine groรe Modellarchitektur zu entwerfen, und viele Mรถglichkeiten, diesen Entwurf zu implementieren. Wie auch immer,
Sie erinnern sich vielleicht an unseren [allgemeinen รberblick รผber ๐ค Transformers](add_new_model#general-overview-of-transformers)
wissen, dass wir ein meinungsfreudiger Haufen sind - die Benutzerfreundlichkeit von ๐ค Transformers hรคngt von konsistenten Designentscheidungen ab. Aus
Erfahrung kรถnnen wir Ihnen ein paar wichtige Dinge รผber das Hinzufรผgen von TensorFlow-Modellen sagen:
- Erfinden Sie das Rad nicht neu! In den meisten Fรคllen gibt es mindestens zwei Referenzimplementierungen, die Sie รผberprรผfen sollten: das
PyTorch-รquivalent des Modells, das Sie implementieren, und andere TensorFlow-Modelle fรผr dieselbe Klasse von Problemen.
- Gute Modellimplementierungen รผberleben den Test der Zeit. Dies geschieht nicht, weil der Code hรผbsch ist, sondern eher
sondern weil der Code klar, einfach zu debuggen und darauf aufzubauen ist. Wenn Sie den Maintainern das Leben mit Ihrer
TensorFlow-Implementierung leicht machen, indem Sie die gleichen Muster wie in anderen TensorFlow-Modellen nachbilden und die Abweichung
zur PyTorch-Implementierung minimieren, stellen Sie sicher, dass Ihr Beitrag lange Bestand haben wird.
- Bitten Sie um Hilfe, wenn Sie nicht weiterkommen! Das ๐ค Transformers-Team ist da, um zu helfen, und wir haben wahrscheinlich Lรถsungen fรผr die gleichen
Probleme gefunden, vor denen Sie stehen.
Hier finden Sie einen รberblick รผber die Schritte, die zum Hinzufรผgen einer TensorFlow-Modellarchitektur erforderlich sind:
1. Wรคhlen Sie das Modell, das Sie konvertieren mรถchten
2. Bereiten Sie die Transformers-Entwicklungsumgebung vor.
3. (Optional) Verstehen Sie die theoretischen Aspekte und die bestehende Implementierung
4. Implementieren Sie die Modellarchitektur
5. Implementieren Sie Modelltests
6. Reichen Sie den Pull-Antrag ein
7. (Optional) Erstellen Sie Demos und teilen Sie diese mit der Welt
### 1.-3. Bereiten Sie Ihren Modellbeitrag vor
**1. Wรคhlen Sie das Modell, das Sie konvertieren mรถchten**
Beginnen wir mit den Grundlagen: Als erstes mรผssen Sie die Architektur kennen, die Sie konvertieren mรถchten. Wenn Sie
Sie sich nicht auf eine bestimmte Architektur festgelegt haben, ist es eine gute Mรถglichkeit, das ๐ค Transformers-Team um Vorschlรคge zu bitten.
Wir werden Sie zu den wichtigsten Architekturen fรผhren, die auf der TensorFlow-Seite noch fehlen.
Seite fehlen. Wenn das spezifische Modell, das Sie mit TensorFlow verwenden mรถchten, bereits eine Implementierung der TensorFlow-Architektur in
๐ค Transformers, aber es fehlen Gewichte, kรถnnen Sie direkt in den
Abschnitt [Gewichtskonvertierung](#adding-tensorflow-weights-to-hub)
auf dieser Seite.
Der Einfachheit halber wird im Rest dieser Anleitung davon ausgegangen, dass Sie sich entschieden haben, mit der TensorFlow-Version von
*BrandNewBert* (dasselbe Beispiel wie in der [Anleitung](add_new_model), um ein neues Modell von Grund auf hinzuzufรผgen).
<Tip>
Bevor Sie mit der Arbeit an einer TensorFlow-Modellarchitektur beginnen, sollten Sie sich vergewissern, dass es keine laufenden Bemรผhungen in dieser Richtung gibt.
Sie kรถnnen nach `BrandNewBert` auf der
[pull request GitHub page](https://github.com/huggingface/transformers/pulls?q=is%3Apr), um zu bestรคtigen, dass es keine
TensorFlow-bezogene Pull-Anfrage gibt.
</Tip>
**2. Transformers-Entwicklungsumgebung vorbereiten**
Nachdem Sie die Modellarchitektur ausgewรคhlt haben, รถffnen Sie einen PR-Entwurf, um Ihre Absicht zu signalisieren, daran zu arbeiten. Folgen Sie den
Anweisungen, um Ihre Umgebung einzurichten und einen PR-Entwurf zu รถffnen.
1. Forken Sie das [repository](https://github.com/huggingface/transformers), indem Sie auf der Seite des Repositorys auf die Schaltflรคche 'Fork' klicken.
Seite des Repositorys klicken. Dadurch wird eine Kopie des Codes unter Ihrem GitHub-Benutzerkonto erstellt.
2. Klonen Sie Ihren `transformers` Fork auf Ihre lokale Festplatte und fรผgen Sie das Basis-Repository als Remote hinzu:
```bash
git clone https://github.com/[your Github handle]/transformers.git
cd transformers
git remote add upstream https://github.com/huggingface/transformers.git
```
3. Richten Sie eine Entwicklungsumgebung ein, indem Sie z.B. den folgenden Befehl ausfรผhren:
```bash
python -m venv .env
source .env/bin/activate
pip install -e ".[dev]"
```
Abhรคngig von Ihrem Betriebssystem und da die Anzahl der optionalen Abhรคngigkeiten von Transformers wรคchst, kann es sein, dass Sie bei diesem Befehl einen
Fehler mit diesem Befehl erhalten. Wenn das der Fall ist, stellen Sie sicher, dass Sie TensorFlow installieren und dann ausfรผhren:
```bash
pip install -e ".[quality]"
```
**Hinweis:** Sie mรผssen CUDA nicht installiert haben. Es reicht aus, das neue Modell auf der CPU laufen zu lassen.
4. Erstellen Sie eine Verzweigung mit einem beschreibenden Namen von Ihrer Hauptverzweigung
```bash
git checkout -b add_tf_brand_new_bert
```
5. Abrufen und zurรผcksetzen auf die aktuelle Hauptversion
```bash
git fetch upstream
git rebase upstream/main
```
6. Fรผgen Sie eine leere `.py` Datei in `transformers/src/models/brandnewbert/` mit dem Namen `modeling_tf_brandnewbert.py` hinzu. Dies wird
Ihre TensorFlow-Modelldatei sein.
7. รbertragen Sie die รnderungen auf Ihr Konto mit:
```bash
git add .
git commit -m "initial commit"
git push -u origin add_tf_brand_new_bert
```
8. Wenn Sie zufrieden sind, gehen Sie auf die Webseite Ihrer Abspaltung auf GitHub. Klicken Sie auf "Pull request". Stellen Sie sicher, dass Sie das
GitHub-Handle einiger Mitglieder des Hugging Face-Teams als Reviewer hinzuzufรผgen, damit das Hugging Face-Team รผber zukรผnftige รnderungen informiert wird.
zukรผnftige รnderungen benachrichtigt wird.
9. รndern Sie den PR in einen Entwurf, indem Sie auf der rechten Seite der GitHub-Pull-Request-Webseite auf "In Entwurf umwandeln" klicken.
Jetzt haben Sie eine Entwicklungsumgebung eingerichtet, um *BrandNewBert* nach TensorFlow in ๐ค Transformers zu portieren.
**3. (Optional) Verstehen Sie die theoretischen Aspekte und die bestehende Implementierung**
Sie sollten sich etwas Zeit nehmen, um die Arbeit von *BrandNewBert* zu lesen, falls eine solche Beschreibung existiert. Mรถglicherweise gibt es groรe
Abschnitte des Papiers, die schwer zu verstehen sind. Wenn das der Fall ist, ist das in Ordnung - machen Sie sich keine Sorgen! Das Ziel ist
ist es nicht, ein tiefes theoretisches Verstรคndnis des Papiers zu erlangen, sondern die notwendigen Informationen zu extrahieren, um
das Modell mit Hilfe von TensorFlow effektiv in ๐ค Transformers neu zu implementieren. Das heiรt, Sie mรผssen nicht zu viel Zeit auf die
viel Zeit auf die theoretischen Aspekte verwenden, sondern sich lieber auf die praktischen Aspekte konzentrieren, nรคmlich auf die bestehende Modelldokumentation
Seite (z.B. [model docs for BERT](model_doc/bert)).
Nachdem Sie die Grundlagen der Modelle, die Sie implementieren wollen, verstanden haben, ist es wichtig, die bestehende
Implementierung zu verstehen. Dies ist eine gute Gelegenheit, sich zu vergewissern, dass eine funktionierende Implementierung mit Ihren Erwartungen an das
Modell entspricht, und um technische Herausforderungen auf der TensorFlow-Seite vorauszusehen.
Es ist ganz natรผrlich, dass Sie sich von der Menge an Informationen, die Sie gerade aufgesogen haben, รผberwรคltigt fรผhlen. Es ist
Es ist definitiv nicht erforderlich, dass Sie in dieser Phase alle Facetten des Modells verstehen. Dennoch empfehlen wir Ihnen dringend
ermutigen wir Sie, alle dringenden Fragen in unserem [Forum](https://discuss.huggingface.co/) zu klรคren.
### 4. Implementierung des Modells
Jetzt ist es an der Zeit, endlich mit dem Programmieren zu beginnen. Als Ausgangspunkt empfehlen wir die PyTorch-Datei selbst: Kopieren Sie den Inhalt von
modeling_brand_new_bert.py` in `src/transformers/models/brand_new_bert/` nach
modeling_tf_brand_new_bert.py`. Das Ziel dieses Abschnitts ist es, die Datei zu รคndern und die Importstruktur von
๐ค Transformers zu aktualisieren, so dass Sie `TFBrandNewBert` und
`TFBrandNewBert.from_pretrained(model_repo, from_pt=True)` erfolgreich ein funktionierendes TensorFlow *BrandNewBert* Modell lรคdt.
Leider gibt es kein Rezept, um ein PyTorch-Modell in TensorFlow zu konvertieren. Sie kรถnnen jedoch unsere Auswahl an
Tipps befolgen, um den Prozess so reibungslos wie mรถglich zu gestalten:
- Stellen Sie `TF` dem Namen aller Klassen voran (z.B. wird `BrandNewBert` zu `TFBrandNewBert`).
- Die meisten PyTorch-Operationen haben einen direkten TensorFlow-Ersatz. Zum Beispiel entspricht `torch.nn.Linear` der Klasse
`tf.keras.layers.Dense`, `torch.nn.Dropout` entspricht `tf.keras.layers.Dropout`, usw. Wenn Sie sich nicht sicher sind
รผber eine bestimmte Operation nicht sicher sind, kรถnnen Sie die [TensorFlow-Dokumentation](https://www.tensorflow.org/api_docs/python/tf)
oder die [PyTorch-Dokumentation](https://pytorch.org/docs/stable/).
- Suchen Sie nach Mustern in der Codebasis von ๐ค Transformers. Wenn Sie auf eine bestimmte Operation stoรen, fรผr die es keinen direkten Ersatz gibt
Ersatz hat, stehen die Chancen gut, dass jemand anderes bereits das gleiche Problem hatte.
- Behalten Sie standardmรครig die gleichen Variablennamen und die gleiche Struktur wie in PyTorch bei. Dies erleichtert die Fehlersuche, die Verfolgung von
Probleme zu verfolgen und spรคtere Korrekturen vorzunehmen.
- Einige Ebenen haben in jedem Framework unterschiedliche Standardwerte. Ein bemerkenswertes Beispiel ist die Schicht fรผr die Batch-Normalisierung
epsilon (`1e-5` in [PyTorch](https://pytorch.org/docs/stable/generated/torch.nn.BatchNorm2d.html#torch.nn.BatchNorm2d)
und `1e-3` in [TensorFlow](https://www.tensorflow.org/api_docs/python/tf/keras/layers/BatchNormalization)).
Prรผfen Sie die Dokumentation genau!
- Die Variablen `nn.Parameter` von PyTorch mรผssen in der Regel innerhalb von TF Layer's `build()` initialisiert werden. Siehe das folgende
Beispiel: [PyTorch](https://github.com/huggingface/transformers/blob/655f72a6896c0533b1bdee519ed65a059c2425ac/src/transformers/models/vit_mae/modeling_vit_mae.py#L212) /
[TensorFlow](https://github.com/huggingface/transformers/blob/655f72a6896c0533b1bdee519ed65a059c2425ac/src/transformers/models/vit_mae/modeling_tf_vit_mae.py#L220)
- Wenn das PyTorch-Modell ein `#copied from ...` am Anfang einer Funktion hat, stehen die Chancen gut, dass Ihr TensorFlow-Modell diese Funktion auch
diese Funktion von der Architektur ausleihen kann, von der sie kopiert wurde, vorausgesetzt, es hat eine TensorFlow-Architektur.
- Die korrekte Zuweisung des Attributs `name` in TensorFlow-Funktionen ist entscheidend, um das `from_pt=True` Gewicht zu erreichen
Cross-Loading. Name" ist fast immer der Name der entsprechenden Variablen im PyTorch-Code. Wenn `name` nicht
nicht richtig gesetzt ist, sehen Sie dies in der Fehlermeldung beim Laden der Modellgewichte.
- Die Logik der Basismodellklasse, `BrandNewBertModel`, befindet sich in `TFBrandNewBertMainLayer`, einer Keras
Schicht-Unterklasse ([Beispiel](https://github.com/huggingface/transformers/blob/4fd32a1f499e45f009c2c0dea4d81c321cba7e02/src/transformers/models/bert/modeling_tf_bert.py#L719)).
TFBrandNewBertModel" ist lediglich ein Wrapper fรผr diese Schicht.
- Keras-Modelle mรผssen erstellt werden, um die vorher trainierten Gewichte zu laden. Aus diesem Grund muss `TFBrandNewBertPreTrainedModel`
ein Beispiel fรผr die Eingaben in das Modell enthalten, die `dummy_inputs`
([Beispiel](https://github.com/huggingface/transformers/blob/4fd32a1f499e45f009c2c0dea4d81c321cba7e02/src/transformers/models/bert/modeling_tf_bert.py#L916)).
- Wenn Sie nicht weiterkommen, fragen Sie nach Hilfe - wir sind fรผr Sie da! ๐ค
Neben der Modelldatei selbst mรผssen Sie auch die Verweise auf die Modellklassen und die zugehรถrigen
Dokumentationsseiten hinzufรผgen. Sie kรถnnen diesen Teil ganz nach den Mustern in anderen PRs erledigen
([Beispiel](https://github.com/huggingface/transformers/pull/18020/files)). Hier ist eine Liste der erforderlichen manuellen
รnderungen:
- Fรผgen Sie alle รถffentlichen Klassen von *BrandNewBert* in `src/transformers/__init__.py` ein.
- Fรผgen Sie *BrandNewBert* Klassen zu den entsprechenden Auto Klassen in `src/transformers/models/auto/modeling_tf_auto.py` hinzu.
- Fรผgen Sie die *BrandNewBert* zugehรถrigen Klassen fรผr trรคges Laden in `src/transformers/utils/dummy_tf_objects.py` hinzu.
- Aktualisieren Sie die Importstrukturen fรผr die รถffentlichen Klassen in `src/transformers/models/brand_new_bert/__init__.py`.
- Fรผgen Sie die Dokumentationszeiger auf die รถffentlichen Methoden von *BrandNewBert* in `docs/source/de/model_doc/brand_new_bert.md` hinzu.
- Fรผgen Sie sich selbst zur Liste der Mitwirkenden an *BrandNewBert* in `docs/source/de/model_doc/brand_new_bert.md` hinzu.
- Fรผgen Sie schlieรlich ein grรผnes Hรคkchen โ
in der TensorFlow-Spalte von *BrandNewBert* in `docs/source/de/index.md` hinzu.
Wenn Sie mit Ihrer Implementierung zufrieden sind, fรผhren Sie die folgende Checkliste aus, um zu bestรคtigen, dass Ihre Modellarchitektur
fertig ist:
1. Alle Schichten, die sich zur Trainingszeit anders verhalten (z.B. Dropout), werden mit einem `Training` Argument aufgerufen, das
von den Top-Level-Klassen weitergegeben wird
2. Sie haben `#copied from ...` verwendet, wann immer es mรถglich war.
3. Die Funktion `TFBrandNewBertMainLayer` und alle Klassen, die sie verwenden, haben ihre Funktion `call` mit `@unpack_inputs` dekoriert
4. TFBrandNewBertMainLayer` ist mit `@keras_serializable` dekoriert
5. Ein TensorFlow-Modell kann aus PyTorch-Gewichten mit `TFBrandNewBert.from_pretrained(model_repo, from_pt=True)` geladen werden.
6. Sie kรถnnen das TensorFlow Modell mit dem erwarteten Eingabeformat aufrufen
### 5. Modell-Tests hinzufรผgen
Hurra, Sie haben ein TensorFlow-Modell implementiert! Jetzt ist es an der Zeit, Tests hinzuzufรผgen, um sicherzustellen, dass sich Ihr Modell wie erwartet verhรคlt.
erwartet. Wie im vorigen Abschnitt schlagen wir vor, dass Sie zunรคchst die Datei `test_modeling_brand_new_bert.py` in
`tests/models/brand_new_bert/` in die Datei `test_modeling_tf_brand_new_bert.py` zu kopieren und dann die notwendigen
TensorFlow-Ersetzungen vornehmen. Fรผr den Moment sollten Sie in allen Aufrufen von `.from_pretrained()` das Flag `from_pt=True` verwenden, um die
die vorhandenen PyTorch-Gewichte zu laden.
Wenn Sie damit fertig sind, kommt der Moment der Wahrheit: Fรผhren Sie die Tests durch! ๐ฌ
```bash
NVIDIA_TF32_OVERRIDE=0 RUN_SLOW=1 RUN_PT_TF_CROSS_TESTS=1 \
py.test -vv tests/models/brand_new_bert/test_modeling_tf_brand_new_bert.py
```
Das wahrscheinlichste Ergebnis ist, dass Sie eine Reihe von Fehlern sehen werden. Machen Sie sich keine Sorgen, das ist zu erwarten! Das Debuggen von ML-Modellen ist
notorisch schwierig, und der Schlรผssel zum Erfolg ist Geduld (und `breakpoint()`). Nach unserer Erfahrung sind die schwierigsten
Probleme aus subtilen Unstimmigkeiten zwischen ML-Frameworks, zu denen wir am Ende dieses Leitfadens ein paar Hinweise geben.
In anderen Fรคllen kann es sein, dass ein allgemeiner Test nicht direkt auf Ihr Modell anwendbar ist; in diesem Fall empfehlen wir eine รberschreibung
auf der Ebene der Modelltestklasse. Zรถgern Sie nicht, in Ihrem Entwurf einer Pull-Anfrage um Hilfe zu bitten, wenn
Sie nicht weiterkommen.
Wenn alle Tests erfolgreich waren, kรถnnen Sie Ihr Modell in die ๐ค Transformers-Bibliothek aufnehmen! ๐
### 6.-7. Stellen Sie sicher, dass jeder Ihr Modell verwenden kann
**6. Reichen Sie den Pull Request ein**
Sobald Sie mit der Implementierung und den Tests fertig sind, ist es an der Zeit, eine Pull-Anfrage einzureichen. Bevor Sie Ihren Code einreichen,
fรผhren Sie unser Dienstprogramm zur Codeformatierung, `make fixup` ๐ช, aus. Damit werden automatisch alle Formatierungsfehler behoben, die dazu fรผhren wรผrden, dass
unsere automatischen Prรผfungen fehlschlagen wรผrden.
Nun ist es an der Zeit, Ihren Entwurf einer Pull-Anfrage in eine echte Pull-Anfrage umzuwandeln. Klicken Sie dazu auf die Schaltflรคche "Bereit fรผr
Review" und fรผgen Sie Joao (`@gante`) und Matt (`@Rocketknight1`) als Reviewer hinzu. Eine Modell-Pull-Anfrage benรถtigt
mindestens 3 Reviewer, aber sie werden sich darum kรผmmern, geeignete zusรคtzliche Reviewer fรผr Ihr Modell zu finden.
Nachdem alle Gutachter mit dem Stand Ihres PR zufrieden sind, entfernen Sie als letzten Aktionspunkt das Flag `from_pt=True` in
.from_pretrained()-Aufrufen zu entfernen. Da es keine TensorFlow-Gewichte gibt, mรผssen Sie sie hinzufรผgen! Lesen Sie den Abschnitt
unten, um zu erfahren, wie Sie dies tun kรถnnen.
Wenn schlieรlich die TensorFlow-Gewichte zusammengefรผhrt werden, Sie mindestens 3 Genehmigungen von Prรผfern haben und alle CI-Checks grรผn sind
grรผn sind, รผberprรผfen Sie die Tests ein letztes Mal lokal
```bash
NVIDIA_TF32_OVERRIDE=0 RUN_SLOW=1 RUN_PT_TF_CROSS_TESTS=1 \
py.test -vv tests/models/brand_new_bert/test_modeling_tf_brand_new_bert.py
```
und wir werden Ihren PR zusammenfรผhren! Herzlichen Glรผckwunsch zu dem Meilenstein ๐.
**7. (Optional) Erstellen Sie Demos und teilen Sie sie mit der Welt**
Eine der schwierigsten Aufgaben bei Open-Source ist die Entdeckung. Wie kรถnnen die anderen Benutzer von der Existenz Ihres
fabelhaften TensorFlow-Beitrags erfahren? Mit der richtigen Kommunikation, natรผrlich! ๐ฃ
Es gibt vor allem zwei Mรถglichkeiten, Ihr Modell mit der Community zu teilen:
- Erstellen Sie Demos. Dazu gehรถren Gradio-Demos, Notebooks und andere unterhaltsame Mรถglichkeiten, Ihr Modell vorzufรผhren. Wir raten Ihnen
ermutigen Sie, ein Notizbuch zu unseren [community-driven demos](https://huggingface.co/docs/transformers/community) hinzuzufรผgen.
- Teilen Sie Geschichten in sozialen Medien wie Twitter und LinkedIn. Sie sollten stolz auf Ihre Arbeit sein und sie mit der
Ihre Leistung mit der Community teilen - Ihr Modell kann nun von Tausenden von Ingenieuren und Forschern auf der ganzen Welt genutzt werden
der Welt genutzt werden ๐! Wir werden Ihre Beitrรคge gerne retweeten und Ihnen helfen, Ihre Arbeit mit der Community zu teilen.
## Hinzufรผgen von TensorFlow-Gewichten zum ๐ค Hub
Unter der Annahme, dass die TensorFlow-Modellarchitektur in ๐ค Transformers verfรผgbar ist, ist die Umwandlung von PyTorch-Gewichten in
TensorFlow-Gewichte ist ein Kinderspiel!
Hier sehen Sie, wie es geht:
1. Stellen Sie sicher, dass Sie in Ihrem Terminal bei Ihrem Hugging Face Konto angemeldet sind. Sie kรถnnen sich mit dem folgenden Befehl anmelden
`huggingface-cli login` (Ihre Zugangstoken finden Sie [hier](https://huggingface.co/settings/tokens))
2. Fรผhren Sie `transformers-cli pt-to-tf --model-name foo/bar` aus, wobei `foo/bar` der Name des Modell-Repositorys ist
ist, das die PyTorch-Gewichte enthรคlt, die Sie konvertieren mรถchten.
3. Markieren Sie `@joaogante` und `@Rocketknight1` in dem ๐ค Hub PR, den der obige Befehl gerade erstellt hat
Das war's! ๐
## Fehlersuche in verschiedenen ML-Frameworks ๐
Irgendwann, wenn Sie eine neue Architektur hinzufรผgen oder TensorFlow-Gewichte fรผr eine bestehende Architektur erstellen, werden Sie
stoรen Sie vielleicht auf Fehler, die sich รผber Unstimmigkeiten zwischen PyTorch und TensorFlow beschweren. Sie kรถnnten sich sogar dazu entschlieรen, den
Modellarchitektur-Code fรผr die beiden Frameworks zu รถffnen, und stellen fest, dass sie identisch aussehen. Was ist denn da los? ๐ค
Lassen Sie uns zunรคchst darรผber sprechen, warum es wichtig ist, diese Diskrepanzen zu verstehen. Viele Community-Mitglieder werden ๐ค
Transformers-Modelle und vertrauen darauf, dass sich unsere Modelle wie erwartet verhalten. Wenn es eine groรe Diskrepanz gibt
zwischen den beiden Frameworks auftritt, bedeutet dies, dass das Modell nicht der Referenzimplementierung fรผr mindestens eines der Frameworks folgt.
der Frameworks folgt. Dies kann zu stillen Fehlern fรผhren, bei denen das Modell zwar lรคuft, aber eine schlechte Leistung aufweist. Dies ist
wohl schlimmer als ein Modell, das รผberhaupt nicht lรคuft! Aus diesem Grund streben wir an, dass die Abweichung zwischen den Frameworks kleiner als
1e-5" in allen Phasen des Modells.
Wie bei anderen numerischen Problemen auch, steckt der Teufel im Detail. Und wie bei jedem detailorientierten Handwerk ist die geheime
Zutat hier Geduld. Hier ist unser Vorschlag fรผr den Arbeitsablauf, wenn Sie auf diese Art von Problemen stoรen:
1. Lokalisieren Sie die Quelle der Abweichungen. Das Modell, das Sie konvertieren, hat wahrscheinlich bis zu einem gewissen Punkt nahezu identische innere Variablen.
bestimmten Punkt. Platzieren Sie `Breakpoint()`-Anweisungen in den Architekturen der beiden Frameworks und vergleichen Sie die Werte der
numerischen Variablen von oben nach unten, bis Sie die Quelle der Probleme gefunden haben.
2. Nachdem Sie nun die Ursache des Problems gefunden haben, setzen Sie sich mit dem ๐ค Transformers-Team in Verbindung. Es ist mรถglich
dass wir ein รคhnliches Problem schon einmal gesehen haben und umgehend eine Lรถsung anbieten kรถnnen. Als Ausweichmรถglichkeit kรถnnen Sie beliebte Seiten
wie StackOverflow und GitHub-Probleme.
3. Wenn keine Lรถsung in Sicht ist, bedeutet das, dass Sie tiefer gehen mรผssen. Die gute Nachricht ist, dass Sie das Problem gefunden haben.
Problem ausfindig gemacht haben, so dass Sie sich auf die problematische Anweisung konzentrieren und den Rest des Modells ausblenden kรถnnen! Die schlechte Nachricht ist
dass Sie sich in die Quellimplementierung der besagten Anweisung einarbeiten mรผssen. In manchen Fรคllen finden Sie vielleicht ein
Problem mit einer Referenzimplementierung - verzichten Sie nicht darauf, ein Problem im Upstream-Repository zu รถffnen.
In einigen Fรคllen kรถnnen wir nach Rรผcksprache mit dem ๐ค Transformers-Team zu dem Schluss kommen, dass die Behebung der Abweichung nicht machbar ist.
Wenn die Abweichung in den Ausgabeschichten des Modells sehr klein ist (aber mรถglicherweise groร in den versteckten Zustรคnden), kรถnnen wir
kรถnnten wir beschlieรen, sie zu ignorieren und das Modell zu verteilen. Die oben erwรคhnte CLI `pt-to-tf` hat ein `--max-error`
Flag, um die Fehlermeldung bei der Gewichtskonvertierung zu รผberschreiben.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/transformers_agents.md
|
<!--Copyright 2023 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Transformers Agents
<Tip warning={true}>
Transformers Agents ist eine experimentelle API, die jederzeit geรคndert werden kann. Die von den Agenten zurรผckgegebenen Ergebnisse
zurรผckgegeben werden, kรถnnen variieren, da sich die APIs oder die zugrunde liegenden Modelle รคndern kรถnnen.
</Tip>
Transformers Version v4.29.0, die auf dem Konzept von *Tools* und *Agenten* aufbaut. Sie kรถnnen damit spielen in
[dieses Colab](https://colab.research.google.com/drive/1c7MHD-T1forUPGcC_jlwsIptOzpG3hSj).
Kurz gesagt, es bietet eine API fรผr natรผrliche Sprache auf der Grundlage von Transformers: Wir definieren eine Reihe von kuratierten Tools und entwerfen einen
Agenten, um natรผrliche Sprache zu interpretieren und diese Werkzeuge zu verwenden. Es ist von vornherein erweiterbar; wir haben einige relevante Tools kuratiert,
aber wir werden Ihnen zeigen, wie das System einfach erweitert werden kann, um jedes von der Community entwickelte Tool zu verwenden.
Beginnen wir mit einigen Beispielen dafรผr, was mit dieser neuen API erreicht werden kann. Sie ist besonders leistungsfรคhig, wenn es um
Sie ist besonders leistungsstark, wenn es um multimodale Aufgaben geht. Lassen Sie uns also eine Runde drehen, um Bilder zu erzeugen und Text vorzulesen.
```py
agent.run("Caption the following image", image=image)
```
| **Input** | **Output** |
|-----------------------------------------------------------------------------------------------------------------------------|-----------------------------------|
| <img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/beaver.png" width=200> | A beaver is swimming in the water |
---
```py
agent.run("Read the following text out loud", text=text)
```
| **Input** | **Output** |
|-------------------------------------------------------------------------------------------------------------------------|----------------------------------------------|
| A beaver is swimming in the water | <audio controls><source src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/tts_example.wav" type="audio/wav"> your browser does not support the audio element. </audio>
---
```py
agent.run(
"In the following `document`, where will the TRRF Scientific Advisory Council Meeting take place?",
document=document,
)
```
| **Input** | **Output** |
|-----------------------------------------------------------------------------------------------------------------------------|----------------|
| <img src="https://datasets-server.huggingface.co/assets/hf-internal-testing/example-documents/--/hf-internal-testing--example-documents/test/0/image/image.jpg" width=200> | ballroom foyer |
## Schnellstart
Bevor Sie `agent.run` verwenden kรถnnen, mรผssen Sie einen Agenten instanziieren, der ein groรes Sprachmodell (LLM) ist.
Wir bieten Unterstรผtzung fรผr openAI-Modelle sowie fรผr OpenSource-Alternativen von BigCode und OpenAssistant. Die openAI
Modelle sind leistungsfรคhiger (erfordern aber einen openAI-API-Schlรผssel, kรถnnen also nicht kostenlos verwendet werden); Hugging Face
bietet kostenlosen Zugang zu Endpunkten fรผr BigCode- und OpenAssistant-Modelle.
To start with, please install the `agents` extras in order to install all default dependencies.
```bash
pip install transformers[agents]
```
Um openAI-Modelle zu verwenden, instanziieren Sie einen [`OpenAiAgent`], nachdem Sie die `openai`-Abhรคngigkeit installiert haben:
```bash
pip install openai
```
```py
from transformers import OpenAiAgent
agent = OpenAiAgent(model="text-davinci-003", api_key="<your_api_key>")
```
Um BigCode oder OpenAssistant zu verwenden, melden Sie sich zunรคchst an, um Zugriff auf die Inference API zu erhalten:
```py
from huggingface_hub import login
login("<YOUR_TOKEN>")
```
Dann instanziieren Sie den Agenten
```py
from transformers import HfAgent
# Starcoder
agent = HfAgent("https://api-inference.huggingface.co/models/bigcode/starcoder")
# StarcoderBase
# agent = HfAgent("https://api-inference.huggingface.co/models/bigcode/starcoderbase")
# OpenAssistant
# agent = HfAgent(url_endpoint="https://api-inference.huggingface.co/models/OpenAssistant/oasst-sft-4-pythia-12b-epoch-3.5")
```
Dies geschieht mit der Inferenz-API, die Hugging Face derzeit kostenlos zur Verfรผgung stellt. Wenn Sie Ihren eigenen Inferenz
Endpunkt fรผr dieses Modell (oder einen anderen) haben, kรถnnen Sie die obige URL durch Ihren URL-Endpunkt ersetzen.
<Tip>
StarCoder und OpenAssistant sind kostenlos und leisten bei einfachen Aufgaben bewundernswert gute Arbeit. Allerdings halten die Kontrollpunkte
nicht, wenn es um komplexere Aufforderungen geht. Wenn Sie mit einem solchen Problem konfrontiert sind, empfehlen wir Ihnen, das OpenAI
Modell auszuprobieren, das zwar leider nicht quelloffen ist, aber zur Zeit eine bessere Leistung erbringt.
</Tip>
Sie sind jetzt startklar! Lassen Sie uns in die beiden APIs eintauchen, die Ihnen jetzt zur Verfรผgung stehen.
### Einzelne Ausfรผhrung (run)
Die Methode der einmaligen Ausfรผhrung ist die Verwendung der [`~Agent.run`] Methode des Agenten:
```py
agent.run("Draw me a picture of rivers and lakes.")
```
<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/rivers_and_lakes.png" width=200>
Es wรคhlt automatisch das (oder die) Werkzeug(e) aus, das (die) fรผr die von Ihnen gewรผnschte Aufgabe geeignet ist (sind) und fรผhrt es (sie) entsprechend aus. Es
kann eine oder mehrere Aufgaben in der gleichen Anweisung ausfรผhren (je komplexer Ihre Anweisung ist, desto wahrscheinlicher ist ein
der Agent scheitern).
```py
agent.run("Draw me a picture of the sea then transform the picture to add an island")
```
<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/sea_and_island.png" width=200>
<br/>
Jede [`~Agent.run`] Operation ist unabhรคngig, so dass Sie sie mehrmals hintereinander mit unterschiedlichen Aufgaben ausfรผhren kรถnnen.
Beachten Sie, dass Ihr `Agent` nur ein groรsprachiges Modell ist, so dass kleine Variationen in Ihrer Eingabeaufforderung vรถllig unterschiedliche Ergebnisse liefern kรถnnen.
unterschiedliche Ergebnisse liefern. Es ist wichtig, dass Sie die Aufgabe, die Sie ausfรผhren mรถchten, so genau wie mรถglich erklรคren. Wir gehen noch weiter ins Detail
wie man gute Prompts schreibt [hier](custom_tools#writing-good-user-inputs).
Wenn Sie einen Status รผber Ausfรผhrungszeiten hinweg beibehalten oder dem Agenten Nicht-Text-Objekte รผbergeben mรถchten, kรถnnen Sie dies tun, indem Sie
Variablen, die der Agent verwenden soll. Sie kรถnnten zum Beispiel das erste Bild von Flรผssen und Seen erzeugen,
und das Modell bitten, dieses Bild zu aktualisieren und eine Insel hinzuzufรผgen, indem Sie Folgendes tun:
```python
picture = agent.run("Generate a picture of rivers and lakes.")
updated_picture = agent.run("Transform the image in `picture` to add an island to it.", picture=picture)
```
<Tip>
Dies kann hilfreich sein, wenn das Modell Ihre Anfrage nicht verstehen kann und die Werkzeuge verwechselt. Ein Beispiel wรคre:
```py
agent.run("Draw me the picture of a capybara swimming in the sea")
```
Hier kรถnnte das Modell auf zwei Arten interpretieren:
- Die Funktion `Text-zu-Bild` erzeugt ein Wasserschwein, das im Meer schwimmt.
- Oder Sie lassen das `Text-zu-Bild` ein Wasserschwein erzeugen und verwenden dann das Werkzeug `Bildtransformation`, um es im Meer schwimmen zu lassen.
Falls Sie das erste Szenario erzwingen mรถchten, kรถnnen Sie dies tun, indem Sie die Eingabeaufforderung als Argument รผbergeben:
```py
agent.run("Draw me a picture of the `prompt`", prompt="a capybara swimming in the sea")
```
</Tip>
### Chat-basierte Ausfรผhrung (Chat)
Der Agent verfรผgt auch รผber einen Chat-basierten Ansatz, der die Methode [`~Agent.chat`] verwendet:
```py
agent.chat("Generate a picture of rivers and lakes")
```
<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/rivers_and_lakes.png" width=200>
```py
agent.chat("Transform the picture so that there is a rock in there")
```
<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/rivers_and_lakes_and_beaver.png" width=200>
<br/>
Dies ist ein interessanter Ansatz, wenn Sie den Zustand รผber Anweisungen hinweg beibehalten mรถchten. Er ist besser fรผr Experimente geeignet,
eignet sich aber eher fรผr einzelne Anweisungen als fรผr komplexe Anweisungen (die die [`~Agent.run`]
Methode besser verarbeiten kann).
Diese Methode kann auch Argumente entgegennehmen, wenn Sie Nicht-Text-Typen oder bestimmte Aufforderungen รผbergeben mรถchten.
### โ ๏ธ Fernausfรผhrung
Zu Demonstrationszwecken und damit es mit allen Setups verwendet werden kann, haben wir Remote-Executors fรผr mehrere
der Standard-Tools erstellt, auf die der Agent in dieser Version Zugriff hat. Diese werden erstellt mit
[inference endpoints](https://huggingface.co/inference-endpoints).
Wir haben diese vorerst deaktiviert, aber um zu sehen, wie Sie selbst Remote Executors Tools einrichten kรถnnen,
empfehlen wir die Lektรผre des [custom tool guide](./custom_tools).
### Was passiert hier? Was sind Tools und was sind Agenten?
<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/diagram.png">
#### Agenten
Der "Agent" ist hier ein groรes Sprachmodell, das wir auffordern, Zugang zu einem bestimmten Satz von Tools zu erhalten.
LLMs sind ziemlich gut darin, kleine Codeproben zu erzeugen. Diese API macht sich das zunutze, indem sie das
LLM ein kleines Codebeispiel gibt, das eine Aufgabe mit einer Reihe von Werkzeugen ausfรผhrt. Diese Aufforderung wird dann ergรคnzt durch die
Aufgabe, die Sie Ihrem Agenten geben, und die Beschreibung der Werkzeuge, die Sie ihm geben. Auf diese Weise erhรคlt er Zugriff auf die Dokumentation der
Tools, insbesondere die erwarteten Eingaben und Ausgaben, und kann den entsprechenden Code generieren.
#### Tools
Tools sind sehr einfach: Sie bestehen aus einer einzigen Funktion mit einem Namen und einer Beschreibung. Wir verwenden dann die Beschreibungen dieser Tools
um den Agenten aufzufordern. Anhand der Eingabeaufforderung zeigen wir dem Agenten, wie er die Tools nutzen kann, um das zu tun, was in der
in der Abfrage angefordert wurde.
Dies geschieht mit brandneuen Tools und nicht mit Pipelines, denn der Agent schreibt besseren Code mit sehr atomaren Tools.
Pipelines sind stรคrker refaktorisiert und fassen oft mehrere Aufgaben in einer einzigen zusammen. Tools sind dafรผr gedacht, sich auf
eine einzige, sehr einfache Aufgabe konzentrieren.
#### Code-Ausfรผhrung?!
Dieser Code wird dann mit unserem kleinen Python-Interpreter auf den mit Ihren Tools รผbergebenen Eingaben ausgefรผhrt.
Wir hรถren Sie schon schreien "Willkรผrliche Codeausfรผhrung!", aber lassen Sie uns erklรคren, warum das nicht der Fall ist.
Die einzigen Funktionen, die aufgerufen werden kรถnnen, sind die von Ihnen zur Verfรผgung gestellten Tools und die Druckfunktion, so dass Sie bereits eingeschrรคnkt sind
eingeschrรคnkt, was ausgefรผhrt werden kann. Sie sollten sicher sein, wenn es sich auf die Werkzeuge fรผr das Umarmungsgesicht beschrรคnkt.
Dann lassen wir keine Attributsuche oder Importe zu (die ohnehin nicht benรถtigt werden, um die
Inputs/Outputs an eine kleine Gruppe von Funktionen), so dass alle offensichtlichen Angriffe (und Sie mรผssten den LLM
dazu auffordern, sie auszugeben) kein Problem darstellen sollten. Wenn Sie auf Nummer sicher gehen wollen, kรถnnen Sie die
run()-Methode mit dem zusรคtzlichen Argument return_code=True ausfรผhren. In diesem Fall gibt der Agent nur den auszufรผhrenden Code
zur Ausfรผhrung zurรผck und Sie kรถnnen entscheiden, ob Sie ihn ausfรผhren mรถchten oder nicht.
Die Ausfรผhrung bricht bei jeder Zeile ab, in der versucht wird, eine illegale Operation auszufรผhren, oder wenn ein regulรคrer Python-Fehler
mit dem vom Agenten generierten Code.
### Ein kuratierter Satz von Tools
Wir haben eine Reihe von Tools identifiziert, die solche Agenten unterstรผtzen kรถnnen. Hier ist eine aktualisierte Liste der Tools, die wir integriert haben
in `transformers` integriert haben:
- **Beantwortung von Fragen zu Dokumenten**: Beantworten Sie anhand eines Dokuments (z.B. PDF) im Bildformat eine Frage zu diesem Dokument ([Donut](./model_doc/donut))
- Beantworten von Textfragen**: Geben Sie einen langen Text und eine Frage an, beantworten Sie die Frage im Text ([Flan-T5](./model_doc/flan-t5))
- **Unbedingte Bildunterschriften**: Beschriften Sie das Bild! ([BLIP](./model_doc/blip))
- **Bildfragebeantwortung**: Beantworten Sie bei einem Bild eine Frage zu diesem Bild ([VILT](./model_doc/vilt))
- **Bildsegmentierung**: Geben Sie ein Bild und einen Prompt an und geben Sie die Segmentierungsmaske dieses Prompts aus ([CLIPSeg](./model_doc/clipseg))
- **Sprache in Text**: Geben Sie eine Audioaufnahme einer sprechenden Person an und transkribieren Sie die Sprache in Text ([Whisper](./model_doc/whisper))
- **Text in Sprache**: wandelt Text in Sprache um ([SpeechT5](./model_doc/speecht5))
- **Zero-Shot-Textklassifizierung**: Ermitteln Sie anhand eines Textes und einer Liste von Bezeichnungen, welcher Bezeichnung der Text am ehesten entspricht ([BART](./model_doc/bart))
- **Textzusammenfassung**: fassen Sie einen langen Text in einem oder wenigen Sรคtzen zusammen ([BART](./model_doc/bart))
- **รbersetzung**: รbersetzen des Textes in eine bestimmte Sprache ([NLLB](./model_doc/nllb))
Diese Tools sind in Transformatoren integriert und kรถnnen auch manuell verwendet werden, zum Beispiel:
```py
from transformers import load_tool
tool = load_tool("text-to-speech")
audio = tool("This is a text to speech tool")
```
### Benutzerdefinierte Tools
Wir haben zwar eine Reihe von Tools identifiziert, sind aber der festen รberzeugung, dass der Hauptwert dieser Implementierung darin besteht
die Mรถglichkeit, benutzerdefinierte Tools schnell zu erstellen und weiterzugeben.
Indem Sie den Code eines Tools in einen Hugging Face Space oder ein Modell-Repository stellen, kรถnnen Sie das Tool
direkt mit dem Agenten nutzen. Wir haben ein paar neue Funktionen hinzugefรผgt
**transformers-agnostic** Tools zur [`huggingface-tools` Organisation](https://huggingface.co/huggingface-tools) hinzugefรผgt:
- **Text-Downloader**: zum Herunterladen eines Textes von einer Web-URL
- **Text zu Bild**: erzeugt ein Bild nach einer Eingabeaufforderung und nutzt dabei stabile Diffusion
- **Bildtransformation**: verรคndert ein Bild anhand eines Ausgangsbildes und einer Eingabeaufforderung, unter Ausnutzung der stabilen pix2pix-Diffusion
- **Text zu Video**: Erzeugen eines kleinen Videos nach einer Eingabeaufforderung, unter Verwendung von damo-vilab
Das Text-zu-Bild-Tool, das wir von Anfang an verwendet haben, ist ein Remote-Tool, das sich in
[*huggingface-tools/text-to-image*](https://huggingface.co/spaces/huggingface-tools/text-to-image)! Wir werden
weiterhin solche Tools fรผr diese und andere Organisationen verรถffentlichen, um diese Implementierung weiter zu verbessern.
Die Agenten haben standardmรครig Zugriff auf die Tools, die sich auf [*huggingface-tools*](https://huggingface.co/huggingface-tools) befinden.
Wie Sie Ihre eigenen Tools schreiben und freigeben kรถnnen und wie Sie jedes benutzerdefinierte Tool, das sich auf dem Hub befindet, nutzen kรถnnen, erklรคren wir in [folgender Anleitung](custom_tools).
### Code-Erzeugung
Bisher haben wir gezeigt, wie Sie die Agenten nutzen kรถnnen, um Aktionen fรผr Sie durchzufรผhren. Der Agent generiert jedoch nur Code
den wir dann mit einem sehr eingeschrรคnkten Python-Interpreter ausfรผhren. Falls Sie den generierten Code in einer anderen Umgebung verwenden mรถchten
einer anderen Umgebung verwenden mรถchten, kรถnnen Sie den Agenten auffordern, den Code zusammen mit einer Tooldefinition und genauen Importen zurรผckzugeben.
Zum Beispiel die folgende Anweisung
```python
agent.run("Draw me a picture of rivers and lakes", return_code=True)
```
gibt den folgenden Code zurรผck
```python
from transformers import load_tool
image_generator = load_tool("huggingface-tools/text-to-image")
image = image_generator(prompt="rivers and lakes")
```
die Sie dann selbst รคndern und ausfรผhren kรถnnen.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/_toctree.yml
|
- sections:
- local: index
title: ๐ค Transformers
- local: quicktour
title: Schnellstart
- local: installation
title: Installation
title: Erste Schritte
- sections:
- local: pipeline_tutorial
title: Pipelines fรผr Inferenzen
- local: autoclass_tutorial
title: Laden von vortrainierten Instanzen mit einer AutoClass
- local: preprocessing
title: Vorverarbeiten
- local: training
title: Optimierung eines vortrainierten Modells
- local: run_scripts
title: Trainieren mit einem Skript
- local: accelerate
title: Verteiltes Training mit ๐ค Accelerate
- local: peft
title: Laden und Trainieren von Adaptern mit ๐ค PEFT
- local: model_sharing
title: Ein Modell teilen
- local: transformers_agents
title: Agents
- local: llm_tutorial
title: Generation with LLMs
title: Tutorials
- sections:
- local: add_new_model
title: Wie fรผgt man ein Modell zu ๐ค Transformers hinzu?
- local: add_tensorflow_model
title: Wie konvertiert man ein ๐ค Transformers-Modell in TensorFlow?
- local: add_new_pipeline
title: Wie fรผgt man eine Pipeline zu ๐ค Transformers hinzu?
- local: testing
title: Testen
- local: pr_checks
title: รberprรผfung einer Pull Request
title: Contribute
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/add_new_model.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Wie kann ich ein Modell zu ๐ค Transformers hinzufรผgen?
Die ๐ค Transformers-Bibliothek ist dank der Beitrรคge der Community oft in der Lage, neue Modelle anzubieten. Aber das kann ein anspruchsvolles Projekt sein und erfordert eine eingehende Kenntnis der ๐ค Transformers-Bibliothek und des zu implementierenden Modells. Bei Hugging Face versuchen wir, mehr Mitgliedern der Community die Mรถglichkeit zu geben, aktiv Modelle hinzuzufรผgen, und wir haben diese Anleitung zusammengestellt, die Sie durch den Prozess des Hinzufรผgens eines PyTorch-Modells fรผhrt (stellen Sie sicher, dass Sie [PyTorch installiert haben](https://pytorch.org/get-started/locally/)).
<Tip>
Wenn Sie daran interessiert sind, ein TensorFlow-Modell zu implementieren, werfen Sie einen Blick in die Anleitung [How to convert a ๐ค Transformers model to TensorFlow](add_tensorflow_model)!
</Tip>
Auf dem Weg dorthin, werden Sie:
- Einblicke in bewรคhrte Open-Source-Verfahren erhalten
- die Konstruktionsprinzipien hinter einer der beliebtesten Deep-Learning-Bibliotheken verstehen
- lernen Sie, wie Sie groรe Modelle effizient testen kรถnnen
- lernen Sie, wie Sie Python-Hilfsprogramme wie `black`, `ruff` und `make fix-copies` integrieren, um sauberen und lesbaren Code zu gewรคhrleisten
Ein Mitglied des Hugging Face-Teams wird Ihnen dabei zur Seite stehen, damit Sie nicht alleine sind. ๐ค โค๏ธ
Um loszulegen, รถffnen Sie eine [New model addition](https://github.com/huggingface/transformers/issues/new?assignees=&labels=New+model&template=new-model-addition.yml) Ausgabe fรผr das Modell, das Sie in ๐ค Transformers sehen mรถchten. Wenn Sie nicht besonders wรคhlerisch sind, wenn es darum geht, ein bestimmtes Modell beizusteuern, kรถnnen Sie nach dem [New model label](https://github.com/huggingface/transformers/labels/New%20model) filtern, um zu sehen, ob es noch unbeanspruchte Modellanfragen gibt, und daran arbeiten.
Sobald Sie eine neue Modellanfrage erรถffnet haben, sollten Sie sich zunรคchst mit ๐ค Transformers vertraut machen, falls Sie das noch nicht sind!
## Allgemeiner รberblick รผber ๐ค Transformers
Zunรคchst sollten Sie sich einen allgemeinen รberblick รผber ๐ค Transformers verschaffen. ๐ค Transformers ist eine sehr meinungsfreudige Bibliothek, es ist also mรถglich, dass
Es besteht also die Mรถglichkeit, dass Sie mit einigen der Philosophien oder Designentscheidungen der Bibliothek nicht einverstanden sind. Aus unserer Erfahrung heraus haben wir jedoch
dass die grundlegenden Designentscheidungen und Philosophien der Bibliothek entscheidend sind, um ๐ค Transformers effizient zu skalieren.
Transformatoren zu skalieren und gleichzeitig die Wartungskosten auf einem vernรผnftigen Niveau zu halten.
Ein guter erster Ansatzpunkt, um die Bibliothek besser zu verstehen, ist die Lektรผre der [Dokumentation unserer Philosophie](Philosophie). Als Ergebnis unserer Arbeitsweise gibt es einige Entscheidungen, die wir versuchen, auf alle Modelle anzuwenden:
- Komposition wird im Allgemeinen gegenรผber Abstraktion bevorzugt
- Die Duplizierung von Code ist nicht immer schlecht, wenn sie die Lesbarkeit oder Zugรคnglichkeit eines Modells stark verbessert
- Modelldateien sind so in sich geschlossen wie mรถglich, so dass Sie, wenn Sie den Code eines bestimmten Modells lesen, idealerweise nur
in die entsprechende Datei `modeling_....py` schauen mรผssen.
Unserer Meinung nach ist der Code der Bibliothek nicht nur ein Mittel, um ein Produkt bereitzustellen, *z.B.* die Mรถglichkeit, BERT fรผr
Inferenz zu verwenden, sondern auch als das Produkt selbst, das wir verbessern wollen. Wenn Sie also ein Modell hinzufรผgen, ist der Benutzer nicht nur die
Person, die Ihr Modell verwenden wird, sondern auch jeder, der Ihren Code liest, zu verstehen versucht und ihn mรถglicherweise verbessert.
Lassen Sie uns daher ein wenig tiefer in das allgemeine Design der Bibliothek einsteigen.
### รberblick รผber die Modelle
Um ein Modell erfolgreich hinzuzufรผgen, ist es wichtig, die Interaktion zwischen Ihrem Modell und seiner Konfiguration zu verstehen,
[`PreTrainedModel`] und [`PretrainedConfig`]. Als Beispiel werden wir
das Modell, das zu ๐ค Transformers hinzugefรผgt werden soll, `BrandNewBert` nennen.
Schauen wir uns das mal an:
<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers_overview.png"/>
Wie Sie sehen, machen wir in ๐ค Transformers von der Vererbung Gebrauch, aber wir beschrรคnken die Abstraktionsebene auf ein absolutes Minimum.
Minimum. Es gibt nie mehr als zwei Abstraktionsebenen fรผr ein Modell in der Bibliothek. `BrandNewBertModel`
erbt von `BrandNewBertPreTrainedModel`, das wiederum von [`PreTrainedModel`] erbt und
das war's. In der Regel wollen wir sicherstellen, dass ein neues Modell nur von
[`PreTrainedModel`] abhรคngt. Die wichtigen Funktionalitรคten, die jedem neuen Modell automatisch zur Verfรผgung gestellt werden, sind
Modell automatisch bereitgestellt werden, sind [`~PreTrainedModel.from_pretrained`] und
[`~PreTrainedModel.save_pretrained`], die fรผr die Serialisierung und Deserialisierung verwendet werden. Alle
anderen wichtigen Funktionalitรคten, wie `BrandNewBertModel.forward` sollten vollstรคndig in der neuen
Skript `modeling_brand_new_bert.py` definiert werden. Als nรคchstes wollen wir sicherstellen, dass ein Modell mit einer bestimmten Kopfebene, wie z.B.
`BrandNewBertForMaskedLM` nicht von `BrandNewBertModel` erbt, sondern `BrandNewBertModel` verwendet
als Komponente, die im Forward Pass aufgerufen werden kann, um die Abstraktionsebene niedrig zu halten. Jedes neue Modell erfordert eine
Konfigurationsklasse, genannt `BrandNewBertConfig`. Diese Konfiguration wird immer als ein Attribut in
[PreTrainedModel] gespeichert und kann daher รผber das Attribut `config` fรผr alle Klassen aufgerufen werden
die von `BrandNewBertPreTrainedModel` erben:
```python
model = BrandNewBertModel.from_pretrained("brandy/brand_new_bert")
model.config # model has access to its config
```
รhnlich wie das Modell erbt die Konfiguration grundlegende Serialisierungs- und Deserialisierungsfunktionalitรคten von
[`PretrainedConfig`]. Beachten Sie, dass die Konfiguration und das Modell immer in zwei verschiedene Formate serialisiert werden
unterschiedliche Formate serialisiert werden - das Modell in eine *pytorch_model.bin* Datei und die Konfiguration in eine *config.json* Datei. Aufruf von
[~PreTrainedModel.save_pretrained`] wird automatisch
[~PretrainedConfig.save_pretrained`] auf, so dass sowohl das Modell als auch die Konfiguration gespeichert werden.
### Code-Stil
Wenn Sie Ihr neues Modell kodieren, sollten Sie daran denken, dass Transformers eine Bibliothek mit vielen Meinungen ist und dass wir selbst ein paar Macken haben
wie der Code geschrieben werden sollte :-)
1. Der Vorwรคrtsdurchlauf Ihres Modells sollte vollstรคndig in die Modellierungsdatei geschrieben werden und dabei vรถllig unabhรคngig von anderen
Modellen in der Bibliothek. Wenn Sie einen Block aus einem anderen Modell wiederverwenden mรถchten, kopieren Sie den Code und fรผgen ihn mit einem
`# Kopiert von` ein (siehe [hier](https://github.com/huggingface/transformers/blob/v4.17.0/src/transformers/models/roberta/modeling_roberta.py#L160)
fรผr ein gutes Beispiel und [hier](pr_checks#check-copies) fรผr weitere Dokumentation zu Copied from).
2. Der Code sollte vollstรคndig verstรคndlich sein, auch fรผr einen Nicht-Muttersprachler. Das heiรt, Sie sollten
beschreibende Variablennamen wรคhlen und Abkรผrzungen vermeiden. Ein Beispiel: `activation` ist `act` vorzuziehen.
Von Variablennamen mit nur einem Buchstaben wird dringend abgeraten, es sei denn, es handelt sich um einen Index in einer for-Schleife.
3. Generell ziehen wir lรคngeren expliziten Code einem kurzen magischen Code vor.
4. Vermeiden Sie die Unterklassifizierung von `nn.Sequential` in PyTorch, sondern unterklassifizieren Sie `nn.Module` und schreiben Sie den Vorwรคrtspass, so dass jeder
so dass jeder, der Ihren Code verwendet, ihn schnell debuggen kann, indem er Druckanweisungen oder Haltepunkte hinzufรผgt.
5. Ihre Funktionssignatur sollte mit einer Typ-Annotation versehen sein. Im รbrigen sind gute Variablennamen viel lesbarer und verstรคndlicher
verstรคndlicher als Typ-Anmerkungen.
### รbersicht der Tokenizer
Noch nicht ganz fertig :-( Dieser Abschnitt wird bald hinzugefรผgt!
## Schritt-fรผr-Schritt-Rezept zum Hinzufรผgen eines Modells zu ๐ค Transformers
Jeder hat andere Vorlieben, was die Portierung eines Modells angeht. Daher kann es sehr hilfreich sein, wenn Sie sich Zusammenfassungen ansehen
wie andere Mitwirkende Modelle auf Hugging Face portiert haben. Hier ist eine Liste von Blogbeitrรคgen aus der Community, wie man ein Modell portiert:
1. [Portierung eines GPT2-Modells](https://medium.com/huggingface/from-tensorflow-to-pytorch-265f40ef2a28) von [Thomas](https://huggingface.co/thomwolf)
2. [Portierung des WMT19 MT-Modells](https://huggingface.co/blog/porting-fsmt) von [Stas](https://huggingface.co/stas)
Aus Erfahrung kรถnnen wir Ihnen sagen, dass die wichtigsten Dinge, die Sie beim Hinzufรผgen eines Modells beachten mรผssen, sind:
- Erfinden Sie das Rad nicht neu! Die meisten Teile des Codes, den Sie fรผr das neue ๐ค Transformers-Modell hinzufรผgen werden, existieren bereits
irgendwo in ๐ค Transformers. Nehmen Sie sich etwas Zeit, um รคhnliche, bereits vorhandene Modelle und Tokenizer zu finden, die Sie kopieren kรถnnen
von. [grep](https://www.gnu.org/software/grep/) und [rg](https://github.com/BurntSushi/ripgrep) sind Ihre
Freunde. Beachten Sie, dass es sehr gut mรถglich ist, dass der Tokenizer Ihres Modells auf einer Modellimplementierung basiert und
und der Modellierungscode Ihres Modells auf einer anderen. *Z.B.* Der Modellierungscode von FSMT basiert auf BART, wรคhrend der Tokenizer-Code von FSMT
auf XLM basiert.
- Es handelt sich eher um eine technische als um eine wissenschaftliche Herausforderung. Sie sollten mehr Zeit auf die Schaffung einer
eine effiziente Debugging-Umgebung zu schaffen, als zu versuchen, alle theoretischen Aspekte des Modells in dem Papier zu verstehen.
- Bitten Sie um Hilfe, wenn Sie nicht weiterkommen! Modelle sind der Kernbestandteil von ๐ค Transformers, so dass wir bei Hugging Face mehr als
mehr als glรผcklich, Ihnen bei jedem Schritt zu helfen, um Ihr Modell hinzuzufรผgen. Zรถgern Sie nicht zu fragen, wenn Sie merken, dass Sie nicht weiterkommen.
Fortschritte machen.
Im Folgenden versuchen wir, Ihnen ein allgemeines Rezept an die Hand zu geben, das uns bei der Portierung eines Modells auf ๐ค Transformers am nรผtzlichsten erschien.
Die folgende Liste ist eine Zusammenfassung all dessen, was getan werden muss, um ein Modell hinzuzufรผgen und kann von Ihnen als To-Do verwendet werden
Liste verwenden:
โ (Optional) Verstehen der theoretischen Aspekte des Modells<br>
โ Vorbereiten der ๐ค Transformers-Entwicklungsumgebung<br>
โ Debugging-Umgebung des ursprรผnglichen Repositorys eingerichtet<br>
โ Skript erstellt, das den Durchlauf `forward()` unter Verwendung des ursprรผnglichen Repositorys und des Checkpoints erfolgreich durchfรผhrt<br>
โ Erfolgreich das Modellskelett zu ๐ค Transformers hinzugefรผgt<br>
โ Erfolgreiche Umwandlung des ursprรผnglichen Prรผfpunkts in den ๐ค Transformers-Prรผfpunkt<br>
โ Erfolgreich den Durchlauf `forward()` in ๐ค Transformers ausgefรผhrt, der eine identische Ausgabe wie der ursprรผngliche Prรผfpunkt liefert<br>
โ Modell-Tests in ๐ค Transformers abgeschlossen<br>
โ Erfolgreich Tokenizer in ๐ค Transformers hinzugefรผgt<br>
โ End-to-End-Integrationstests ausgefรผhrt<br>
โ Docs fertiggestellt<br>
โ Modellgewichte in den Hub hochgeladen<br>
โ Die Pull-Anfrage eingereicht<br>
โ (Optional) Hinzufรผgen eines Demo-Notizbuchs
Fรผr den Anfang empfehlen wir in der Regel, mit einem guten theoretischen Verstรคndnis von `BrandNewBert` zu beginnen. Wie auch immer,
wenn Sie es vorziehen, die theoretischen Aspekte des Modells *on-the-job* zu verstehen, dann ist es vรถllig in Ordnung, direkt in die
in die Code-Basis von `BrandNewBert` einzutauchen. Diese Option kรถnnte fรผr Sie besser geeignet sein, wenn Ihre technischen Fรคhigkeiten besser sind als
als Ihre theoretischen Fรคhigkeiten, wenn Sie Schwierigkeiten haben, die Arbeit von `BrandNewBert` zu verstehen, oder wenn Sie einfach Spaร am Programmieren
mehr Spaร am Programmieren haben als am Lesen wissenschaftlicher Abhandlungen.
### 1. (Optional) Theoretische Aspekte von BrandNewBert
Sie sollten sich etwas Zeit nehmen, um die Abhandlung von *BrandNewBert* zu lesen, falls eine solche Beschreibung existiert. Mรถglicherweise gibt es groรe
Abschnitte des Papiers, die schwer zu verstehen sind. Wenn das der Fall ist, ist das in Ordnung - machen Sie sich keine Sorgen! Das Ziel ist
ist es nicht, ein tiefes theoretisches Verstรคndnis des Papiers zu erlangen, sondern die notwendigen Informationen zu extrahieren, um
das Modell effektiv in ๐ค Transformers zu implementieren. Das heiรt, Sie mรผssen nicht zu viel Zeit auf die
theoretischen Aspekten verbringen, sondern sich lieber auf die praktischen Aspekte konzentrieren, nรคmlich:
- Welche Art von Modell ist *brand_new_bert*? BERT-รคhnliches Modell nur fรผr den Encoder? GPT2-รคhnliches reines Decoder-Modell? BART-รคhnliches
Encoder-Decoder-Modell? Sehen Sie sich die [model_summary](model_summary) an, wenn Sie mit den Unterschieden zwischen diesen Modellen nicht vertraut sind.
- Was sind die Anwendungen von *brand_new_bert*? Textklassifizierung? Texterzeugung? Seq2Seq-Aufgaben, *z.B.,*
Zusammenfassungen?
- Was ist die neue Eigenschaft des Modells, die es von BERT/GPT-2/BART unterscheidet?
- Welches der bereits existierenden [๐ค Transformers-Modelle](https://huggingface.co/transformers/#contents) ist am รคhnlichsten
รคhnlich wie *brand_new_bert*?
- Welche Art von Tokenizer wird verwendet? Ein Satzteil-Tokenisierer? Ein Wortstรผck-Tokenisierer? Ist es derselbe Tokenisierer, der fรผr
fรผr BERT oder BART?
Nachdem Sie das Gefรผhl haben, einen guten รberblick รผber die Architektur des Modells erhalten zu haben, kรถnnen Sie dem
Hugging Face Team schreiben und Ihre Fragen stellen. Dazu kรถnnen Fragen zur Architektur des Modells gehรถren,
seiner Aufmerksamkeitsebene usw. Wir werden Ihnen gerne weiterhelfen.
### 2. Bereiten Sie als nรคchstes Ihre Umgebung vor
1. Forken Sie das [Repository](https://github.com/huggingface/transformers), indem Sie auf der Seite des Repositorys auf die Schaltflรคche 'Fork' klicken.
Seite des Repositorys klicken. Dadurch wird eine Kopie des Codes unter Ihrem GitHub-Benutzerkonto erstellt.
2. Klonen Sie Ihren `transformers` Fork auf Ihre lokale Festplatte und fรผgen Sie das Basis-Repository als Remote hinzu:
```bash
git clone https://github.com/[your Github handle]/transformers.git
cd transformers
git remote add upstream https://github.com/huggingface/transformers.git
```
3. Richten Sie eine Entwicklungsumgebung ein, indem Sie z.B. den folgenden Befehl ausfรผhren:
```bash
python -m venv .env
source .env/bin/activate
pip install -e ".[dev]"
```
Abhรคngig von Ihrem Betriebssystem und da die Anzahl der optionalen Abhรคngigkeiten von Transformers wรคchst, kann es sein, dass Sie bei diesem Befehl einen
Fehler mit diesem Befehl. Stellen Sie in diesem Fall sicher, dass Sie das Deep Learning Framework, mit dem Sie arbeiten, installieren
(PyTorch, TensorFlow und/oder Flax) und fรผhren Sie es aus:
```bash
pip install -e ".[quality]"
```
was fรผr die meisten Anwendungsfรคlle ausreichend sein sollte. Sie kรถnnen dann zum รผbergeordneten Verzeichnis zurรผckkehren
```bash
cd ..
```
4. Wir empfehlen, die PyTorch-Version von *brand_new_bert* zu Transformers hinzuzufรผgen. Um PyTorch zu installieren, folgen Sie bitte den
Anweisungen auf https://pytorch.org/get-started/locally/.
**Anmerkung:** Sie mรผssen CUDA nicht installiert haben. Es reicht aus, das neue Modell auf der CPU zum Laufen zu bringen.
5. Um *brand_new_bert* zu portieren, benรถtigen Sie auรerdem Zugriff auf das Original-Repository:
```bash
git clone https://github.com/org_that_created_brand_new_bert_org/brand_new_bert.git
cd brand_new_bert
pip install -e .
```
Jetzt haben Sie eine Entwicklungsumgebung eingerichtet, um *brand_new_bert* auf ๐ค Transformers zu portieren.
### 3.-4. Fรผhren Sie einen Pre-Training-Checkpoint mit dem Original-Repository durch
Zunรคchst werden Sie mit dem ursprรผnglichen *brand_new_bert* Repository arbeiten. Oft ist die ursprรผngliche Implementierung sehr
"forschungslastig". Das bedeutet, dass es an Dokumentation mangeln kann und der Code schwer zu verstehen sein kann. Aber das sollte
genau Ihre Motivation sein, *brand_new_bert* neu zu implementieren. Eines unserer Hauptziele bei Hugging Face ist es, *die Menschen dazu zu bringen
auf den Schultern von Giganten zu stehen*, was sich hier sehr gut darin ausdrรผckt, dass wir ein funktionierendes Modell nehmen und es umschreiben, um es so
es so **zugรคnglich, benutzerfreundlich und schรถn** wie mรถglich zu machen. Dies ist die wichtigste Motivation fรผr die Neuimplementierung von
Modelle in ๐ค Transformers umzuwandeln - der Versuch, komplexe neue NLP-Technologie fรผr **jeden** zugรคnglich zu machen.
Sie sollten damit beginnen, indem Sie in das Original-Repository eintauchen.
Die erfolgreiche Ausfรผhrung des offiziellen Pre-Trainingsmodells im Original-Repository ist oft **der schwierigste** Schritt.
Unserer Erfahrung nach ist es sehr wichtig, dass Sie einige Zeit damit verbringen, sich mit der ursprรผnglichen Code-Basis vertraut zu machen. Sie mรผssen
das Folgende herausfinden:
- Wo finden Sie die vortrainierten Gewichte?
- Wie lรคdt man die vorab trainierten Gewichte in das entsprechende Modell?
- Wie kann der Tokenizer unabhรคngig vom Modell ausgefรผhrt werden?
- Verfolgen Sie einen Forward Pass, damit Sie wissen, welche Klassen und Funktionen fรผr einen einfachen Forward Pass erforderlich sind. Normalerweise,
mรผssen Sie nur diese Funktionen reimplementieren.
- Sie mรผssen in der Lage sein, die wichtigen Komponenten des Modells zu finden: Wo befindet sich die Klasse des Modells? Gibt es Unterklassen des Modells,
*z.B.* EncoderModel, DecoderModel? Wo befindet sich die Selbstaufmerksamkeitsschicht? Gibt es mehrere verschiedene Aufmerksamkeitsebenen,
*z.B.* *Selbstaufmerksamkeit*, *Kreuzaufmerksamkeit*...?
- Wie kรถnnen Sie das Modell in der ursprรผnglichen Umgebung des Repo debuggen? Mรผssen Sie *print* Anweisungen hinzufรผgen, kรถnnen Sie
mit einem interaktiven Debugger wie *ipdb* arbeiten oder sollten Sie eine effiziente IDE zum Debuggen des Modells verwenden, wie z.B. PyCharm?
Es ist sehr wichtig, dass Sie, bevor Sie mit der Portierung beginnen, den Code im Original-Repository **effizient** debuggen kรถnnen
Repository kรถnnen! Denken Sie auch daran, dass Sie mit einer Open-Source-Bibliothek arbeiten, also zรถgern Sie nicht, ein Problem oder
oder sogar eine Pull-Anfrage im Original-Repository zu stellen. Die Betreuer dieses Repositorys sind wahrscheinlich sehr froh darรผber
dass jemand in ihren Code schaut!
An diesem Punkt liegt es wirklich an Ihnen, welche Debugging-Umgebung und Strategie Sie zum Debuggen des ursprรผnglichen
Modell zu debuggen. Wir raten dringend davon ab, eine kostspielige GPU-Umgebung einzurichten, sondern arbeiten Sie einfach auf einer CPU, sowohl wenn Sie mit dem
in das ursprรผngliche Repository einzutauchen und auch, wenn Sie beginnen, die ๐ค Transformers-Implementierung des Modells zu schreiben. Nur
ganz am Ende, wenn das Modell bereits erfolgreich auf ๐ค Transformers portiert wurde, sollte man รผberprรผfen, ob das
Modell auch auf der GPU wie erwartet funktioniert.
Im Allgemeinen gibt es zwei mรถgliche Debugging-Umgebungen fรผr die Ausfรผhrung des Originalmodells
- [Jupyter notebooks](https://jupyter.org/) / [google colab](https://colab.research.google.com/notebooks/intro.ipynb)
- Lokale Python-Skripte.
Jupyter-Notebooks haben den Vorteil, dass sie eine zellenweise Ausfรผhrung ermรถglichen, was hilfreich sein kann, um logische Komponenten besser voneinander zu trennen und
logische Komponenten voneinander zu trennen und schnellere Debugging-Zyklen zu haben, da Zwischenergebnisse gespeichert werden kรถnnen. Auรerdem,
Auรerdem lassen sich Notebooks oft leichter mit anderen Mitwirkenden teilen, was sehr hilfreich sein kann, wenn Sie das Hugging Face Team um Hilfe bitten mรถchten.
Face Team um Hilfe bitten. Wenn Sie mit Jupyter-Notizbรผchern vertraut sind, empfehlen wir Ihnen dringend, mit ihnen zu arbeiten.
Der offensichtliche Nachteil von Jupyter-Notizbรผchern ist, dass Sie, wenn Sie nicht daran gewรถhnt sind, mit ihnen zu arbeiten, einige Zeit damit verbringen mรผssen
einige Zeit damit verbringen mรผssen, sich an die neue Programmierumgebung zu gewรถhnen, und dass Sie mรถglicherweise Ihre bekannten Debugging-Tools nicht mehr verwenden kรถnnen
wie z.B. `ipdb` nicht mehr verwenden kรถnnen.
Fรผr jede Codebasis ist es immer ein guter erster Schritt, einen **kleinen** vortrainierten Checkpoint zu laden und in der Lage zu sein, einen
einzelnen Vorwรคrtsdurchlauf mit einem Dummy-Integer-Vektor von Eingabe-IDs als Eingabe zu reproduzieren. Ein solches Skript kรถnnte wie folgt aussehen (in
Pseudocode):
```python
model = BrandNewBertModel.load_pretrained_checkpoint("/path/to/checkpoint/")
input_ids = [0, 4, 5, 2, 3, 7, 9] # vector of input ids
original_output = model.predict(input_ids)
```
Was die Debugging-Strategie anbelangt, so kรถnnen Sie im Allgemeinen aus mehreren Strategien wรคhlen:
- Zerlegen Sie das ursprรผngliche Modell in viele kleine testbare Komponenten und fรผhren Sie fรผr jede dieser Komponenten einen Vorwรคrtsdurchlauf zur
รberprรผfung
- Zerlegen Sie das ursprรผngliche Modell nur in den ursprรผnglichen *Tokenizer* und das ursprรผngliche *Modell*, fรผhren Sie einen Vorwรคrtsdurchlauf fรผr diese Komponenten durch
und verwenden Sie dazwischenliegende Druckanweisungen oder Haltepunkte zur รberprรผfung.
Auch hier bleibt es Ihnen รผberlassen, welche Strategie Sie wรคhlen. Oft ist die eine oder die andere Strategie vorteilhaft, je nach der ursprรผnglichen Codebasis
Basis.
Wenn die ursprรผngliche Codebasis es Ihnen erlaubt, das Modell in kleinere Teilkomponenten zu zerlegen, *z.B.* wenn die ursprรผngliche
Code-Basis problemlos im Eager-Modus ausgefรผhrt werden kann, lohnt es sich in der Regel, dies zu tun. Es gibt einige wichtige Vorteile
am Anfang den schwierigeren Weg zu gehen:
- Wenn Sie spรคter das ursprรผngliche Modell mit der Hugging Face-Implementierung vergleichen, kรถnnen Sie automatisch รผberprรผfen, ob
fรผr jede Komponente einzeln รผberprรผfen, ob die entsprechende Komponente der ๐ค Transformers-Implementierung รผbereinstimmt, anstatt sich auf
anstatt sich auf den visuellen Vergleich รผber Druckanweisungen zu verlassen
- kรถnnen Sie das groรe Problem der Portierung eines Modells in kleinere Probleme der Portierung einzelner Komponenten zerlegen
einzelnen Komponenten zu zerlegen und so Ihre Arbeit besser zu strukturieren
- Die Aufteilung des Modells in logisch sinnvolle Komponenten hilft Ihnen, einen besseren รberblick รผber das Design des Modells zu bekommen
und somit das Modell besser zu verstehen
- In einem spรคteren Stadium helfen Ihnen diese komponentenweisen Tests dabei, sicherzustellen, dass keine Regressionen auftreten, wรคhrend Sie fortfahren
Ihren Code รคndern
[Lysandre's](https://gist.github.com/LysandreJik/db4c948f6b4483960de5cbac598ad4ed) Integrationstests fรผr ELECTRA
gibt ein schรถnes Beispiel dafรผr, wie dies geschehen kann.
Wenn die ursprรผngliche Codebasis jedoch sehr komplex ist oder nur die Ausfรผhrung von Zwischenkomponenten in einem kompilierten Modus erlaubt,
kรถnnte es zu zeitaufwรคndig oder sogar unmรถglich sein, das Modell in kleinere testbare Teilkomponenten zu zerlegen. Ein gutes
Beispiel ist die [T5's MeshTensorFlow](https://github.com/tensorflow/mesh/tree/master/mesh_tensorflow) Bibliothek, die sehr komplex ist
sehr komplex ist und keine einfache Mรถglichkeit bietet, das Modell in seine Unterkomponenten zu zerlegen. Bei solchen Bibliotheken ist man
oft auf die รberprรผfung von Druckanweisungen angewiesen.
Unabhรคngig davon, welche Strategie Sie wรคhlen, ist die empfohlene Vorgehensweise oft die gleiche, nรคmlich dass Sie mit der Fehlersuche in den
die Anfangsebenen zuerst und die Endebenen zuletzt debuggen.
Es wird empfohlen, dass Sie die Ausgaben der folgenden Ebenen abrufen, entweder durch Druckanweisungen oder Unterkomponentenfunktionen
Schichten in der folgenden Reihenfolge abrufen:
1. Rufen Sie die Eingabe-IDs ab, die an das Modell รผbergeben wurden
2. Rufen Sie die Worteinbettungen ab
3. Rufen Sie die Eingabe der ersten Transformer-Schicht ab
4. Rufen Sie die Ausgabe der ersten Transformer-Schicht ab
5. Rufen Sie die Ausgabe der folgenden n - 1 Transformer-Schichten ab
6. Rufen Sie die Ausgabe des gesamten BrandNewBert Modells ab
Die Eingabe-IDs sollten dabei aus einem Array von Ganzzahlen bestehen, *z.B.* `input_ids = [0, 4, 4, 3, 2, 4, 1, 7, 19]`
Die Ausgaben der folgenden Schichten bestehen oft aus mehrdimensionalen Float-Arrays und kรถnnen wie folgt aussehen:
```
[[
[-0.1465, -0.6501, 0.1993, ..., 0.1451, 0.3430, 0.6024],
[-0.4417, -0.5920, 0.3450, ..., -0.3062, 0.6182, 0.7132],
[-0.5009, -0.7122, 0.4548, ..., -0.3662, 0.6091, 0.7648],
...,
[-0.5613, -0.6332, 0.4324, ..., -0.3792, 0.7372, 0.9288],
[-0.5416, -0.6345, 0.4180, ..., -0.3564, 0.6992, 0.9191],
[-0.5334, -0.6403, 0.4271, ..., -0.3339, 0.6533, 0.8694]]],
```
Wir erwarten, dass jedes zu ๐ค Transformers hinzugefรผgte Modell eine Reihe von Integrationstests besteht, was bedeutet, dass das ursprรผngliche
Modell und die neu implementierte Version in ๐ค Transformers exakt dieselbe Ausgabe liefern mรผssen, und zwar mit einer Genauigkeit von 0,001!
Da es normal ist, dass das exakt gleiche Modell, das in verschiedenen Bibliotheken geschrieben wurde, je nach Bibliotheksrahmen eine leicht unterschiedliche Ausgabe liefern kann
eine leicht unterschiedliche Ausgabe liefern kann, akzeptieren wir eine Fehlertoleranz von 1e-3 (0,001). Es reicht nicht aus, wenn das Modell
fast das gleiche Ergebnis liefert, sie mรผssen fast identisch sein. Daher werden Sie sicherlich die Zwischenergebnisse
Zwischenergebnisse der ๐ค Transformers-Version mehrfach mit den Zwischenergebnissen der ursprรผnglichen Implementierung von
*brand_new_bert* vergleichen. In diesem Fall ist eine **effiziente** Debugging-Umgebung des ursprรผnglichen Repositorys absolut
wichtig ist. Hier sind einige Ratschlรคge, um Ihre Debugging-Umgebung so effizient wie mรถglich zu gestalten.
- Finden Sie den besten Weg, um Zwischenergebnisse zu debuggen. Ist das ursprรผngliche Repository in PyTorch geschrieben? Dann sollten Sie
dann sollten Sie sich wahrscheinlich die Zeit nehmen, ein lรคngeres Skript zu schreiben, das das ursprรผngliche Modell in kleinere Unterkomponenten zerlegt, um
Zwischenwerte abzurufen. Ist das ursprรผngliche Repository in Tensorflow 1 geschrieben? Dann mรผssen Sie sich mรถglicherweise auf die
TensorFlow Druckoperationen wie [tf.print](https://www.tensorflow.org/api_docs/python/tf/print) verlassen, um die
Zwischenwerte auszugeben. Ist das ursprรผngliche Repository in Jax geschrieben? Dann stellen Sie sicher, dass das Modell **nicht jitted** ist, wenn
wenn Sie den Vorwรคrtsdurchlauf ausfรผhren, *z.B.* schauen Sie sich [dieser Link](https://github.com/google/jax/issues/196) an.
- Verwenden Sie den kleinsten vortrainierten Prรผfpunkt, den Sie finden kรถnnen. Je kleiner der Prรผfpunkt ist, desto schneller wird Ihr Debugging-Zyklus
wird. Es ist nicht effizient, wenn Ihr vorab trainiertes Modell so groร ist, dass Ihr Vorwรคrtsdurchlauf mehr als 10 Sekunden dauert.
Falls nur sehr groรe Checkpoints verfรผgbar sind, kann es sinnvoller sein, ein Dummy-Modell in der neuen
Umgebung mit zufรคllig initialisierten Gewichten zu erstellen und diese Gewichte zum Vergleich mit der ๐ค Transformers-Version
Ihres Modells
- Vergewissern Sie sich, dass Sie den einfachsten Weg wรคhlen, um einen Forward Pass im ursprรผnglichen Repository aufzurufen. Idealerweise sollten Sie
die Funktion im originalen Repository finden, die **nur** einen einzigen Vorwรคrtspass aufruft, *d.h.* die oft aufgerufen wird
Vorhersagen", "Auswerten", "Vorwรคrts" oder "Aufruf" genannt wird. Sie wollen keine Funktion debuggen, die `forward` aufruft
mehrfach aufruft, *z.B.* um Text zu erzeugen, wie `autoregressive_sample`, `generate`.
- Versuchen Sie, die Tokenisierung vom *Forward*-Pass des Modells zu trennen. Wenn das Original-Repository Beispiele zeigt, bei denen
Sie eine Zeichenkette eingeben mรผssen, dann versuchen Sie herauszufinden, an welcher Stelle im Vorwรคrtsaufruf die Zeichenketteneingabe in Eingabe-IDs geรคndert wird
geรคndert wird und beginnen Sie an dieser Stelle. Das kรถnnte bedeuten, dass Sie mรถglicherweise selbst ein kleines Skript schreiben oder den
Originalcode so รคndern mรผssen, dass Sie die ids direkt eingeben kรถnnen, anstatt eine Zeichenkette einzugeben.
- Vergewissern Sie sich, dass sich das Modell in Ihrem Debugging-Setup **nicht** im Trainingsmodus befindet, der oft dazu fรผhrt, dass das Modell
Dies fรผhrt hรคufig zu zufรคlligen Ergebnissen, da das Modell mehrere Dropout-Schichten enthรคlt. Stellen Sie sicher, dass der Vorwรคrtsdurchlauf in Ihrer Debugging
Umgebung **deterministisch** ist, damit die Dropout-Schichten nicht verwendet werden. Oder verwenden Sie *transformers.utils.set_seed*.
wenn sich die alte und die neue Implementierung im selben Framework befinden.
Im folgenden Abschnitt finden Sie genauere Details/Tipps, wie Sie dies fรผr *brand_new_bert* tun kรถnnen.
### 5.-14. Portierung von BrandNewBert auf ๐ค Transformatoren
Als nรคchstes kรถnnen Sie endlich damit beginnen, neuen Code zu ๐ค Transformers hinzuzufรผgen. Gehen Sie in den Klon Ihres ๐ค Transformers Forks:
```bash
cd transformers
```
In dem speziellen Fall, dass Sie ein Modell hinzufรผgen, dessen Architektur genau mit der Modellarchitektur eines
Modells รผbereinstimmt, mรผssen Sie nur ein Konvertierungsskript hinzufรผgen, wie in [diesem Abschnitt](#write-a-conversion-script) beschrieben.
In diesem Fall kรถnnen Sie einfach die gesamte Modellarchitektur des bereits vorhandenen Modells wiederverwenden.
Andernfalls beginnen wir mit der Erstellung eines neuen Modells. Sie haben hier zwei Mรถglichkeiten:
- `transformers-cli add-new-model-like`, um ein neues Modell wie ein bestehendes hinzuzufรผgen
- `transformers-cli add-new-model`, um ein neues Modell aus unserer Vorlage hinzuzufรผgen (sieht dann aus wie BERT oder Bart, je nachdem, welche Art von Modell Sie wรคhlen)
In beiden Fรคllen werden Sie mit einem Fragebogen aufgefordert, die grundlegenden Informationen zu Ihrem Modell auszufรผllen. Fรผr den zweiten Befehl mรผssen Sie `cookiecutter` installieren, weitere Informationen dazu finden Sie [hier](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model).
**Erรถffnen Sie einen Pull Request auf dem Haupt-Repositorium huggingface/transformers**
Bevor Sie mit der Anpassung des automatisch generierten Codes beginnen, ist es nun an der Zeit, einen "Work in progress (WIP)" Pull
Anfrage, *z.B.* "[WIP] Add *brand_new_bert*", in ๐ค Transformers zu รถffnen, damit Sie und das Hugging Face Team
Seite an Seite an der Integration des Modells in ๐ค Transformers arbeiten kรถnnen.
Sie sollten Folgendes tun:
1. Erstellen Sie eine Verzweigung mit einem beschreibenden Namen von Ihrer Hauptverzweigung
```bash
git checkout -b add_brand_new_bert
```
2. Bestรคtigen Sie den automatisch generierten Code:
```bash
git add .
git commit
```
3. Abrufen und zurรผcksetzen auf die aktuelle Haupt
```bash
git fetch upstream
git rebase upstream/main
```
4. รbertragen Sie die รnderungen auf Ihr Konto mit:
```bash
git push -u origin a-descriptive-name-for-my-changes
```
5. Wenn Sie zufrieden sind, gehen Sie auf die Webseite Ihrer Abspaltung auf GitHub. Klicken Sie auf "Pull request". Stellen Sie sicher, dass Sie das
GitHub-Handle einiger Mitglieder des Hugging Face-Teams als Reviewer hinzuzufรผgen, damit das Hugging Face-Team รผber zukรผnftige รnderungen informiert wird.
zukรผnftige รnderungen benachrichtigt wird.
6. รndern Sie den PR in einen Entwurf, indem Sie auf der rechten Seite der GitHub-Pull-Request-Webseite auf "In Entwurf umwandeln" klicken.
Vergessen Sie im Folgenden nicht, wenn Sie Fortschritte gemacht haben, Ihre Arbeit zu committen und in Ihr Konto zu pushen, damit sie in der Pull-Anfrage erscheint.
damit sie in der Pull-Anfrage angezeigt wird. Auรerdem sollten Sie darauf achten, dass Sie Ihre Arbeit von Zeit zu Zeit mit dem aktuellen main
von Zeit zu Zeit zu aktualisieren, indem Sie dies tun:
```bash
git fetch upstream
git merge upstream/main
```
Generell sollten Sie alle Fragen, die Sie in Bezug auf das Modell oder Ihre Implementierung haben, in Ihrem PR stellen und
in der PR diskutiert/gelรถst werden. Auf diese Weise wird das Hugging Face Team immer benachrichtigt, wenn Sie neuen Code einreichen oder
wenn Sie eine Frage haben. Es ist oft sehr hilfreich, das Hugging Face-Team auf Ihren hinzugefรผgten Code hinzuweisen, damit das Hugging Face-Team Ihr Problem oder Ihre Frage besser verstehen kann.
Face-Team Ihr Problem oder Ihre Frage besser verstehen kann.
Gehen Sie dazu auf die Registerkarte "Geรคnderte Dateien", auf der Sie alle Ihre รnderungen sehen, gehen Sie zu einer Zeile, zu der Sie eine Frage stellen mรถchten
eine Frage stellen mรถchten, und klicken Sie auf das "+"-Symbol, um einen Kommentar hinzuzufรผgen. Wenn eine Frage oder ein Problem gelรถst wurde,
kรถnnen Sie auf die Schaltflรคche "Lรถsen" des erstellten Kommentars klicken.
Auf dieselbe Weise wird das Hugging Face-Team Kommentare รถffnen, wenn es Ihren Code รผberprรผft. Wir empfehlen, die meisten Fragen
auf GitHub in Ihrem PR zu stellen. Fรผr einige sehr allgemeine Fragen, die fรผr die รffentlichkeit nicht sehr nรผtzlich sind, kรถnnen Sie das
Hugging Face Team per Slack oder E-Mail zu stellen.
**5. Passen Sie den Code der generierten Modelle fรผr brand_new_bert** an.
Zunรคchst werden wir uns nur auf das Modell selbst konzentrieren und uns nicht um den Tokenizer kรผmmern. Den gesamten relevanten Code sollten Sie
finden Sie in den generierten Dateien `src/transformers/models/brand_new_bert/modeling_brand_new_bert.py` und
`src/transformers/models/brand_new_bert/configuration_brand_new_bert.py`.
Jetzt kรถnnen Sie endlich mit dem Programmieren beginnen :). Der generierte Code in
`src/transformers/models/brand_new_bert/modeling_brand_new_bert.py` wird entweder die gleiche Architektur wie BERT haben, wenn
wenn es sich um ein reines Encoder-Modell handelt oder BART, wenn es sich um ein Encoder-Decoder-Modell handelt. An diesem Punkt sollten Sie sich daran erinnern, was
was Sie am Anfang รผber die theoretischen Aspekte des Modells gelernt haben: *Wie unterscheidet sich das Modell von BERT oder
BART?*". Implementieren Sie diese รnderungen, was oft bedeutet, dass Sie die *Selbstaufmerksamkeitsschicht*, die Reihenfolge der Normalisierungsschicht usw. รคndern mรผssen.
Schicht usw... Auch hier ist es oft nรผtzlich, sich die รคhnliche Architektur bereits bestehender Modelle in Transformers anzusehen, um ein besseres Gefรผhl dafรผr zu bekommen
ein besseres Gefรผhl dafรผr zu bekommen, wie Ihr Modell implementiert werden sollte.
**Beachten Sie**, dass Sie an diesem Punkt nicht sehr sicher sein mรผssen, dass Ihr Code vรถllig korrekt oder sauber ist. Vielmehr ist es
Sie sollten vielmehr eine erste *unbereinigte*, kopierte Version des ursprรผnglichen Codes in
src/transformers/models/brand_new_bert/modeling_brand_new_bert.py" hinzuzufรผgen, bis Sie das Gefรผhl haben, dass der gesamte notwendige Code
hinzugefรผgt wurde. Unserer Erfahrung nach ist es viel effizienter, schnell eine erste Version des erforderlichen Codes hinzuzufรผgen und
den Code iterativ mit dem Konvertierungsskript zu verbessern/korrigieren, wie im nรคchsten Abschnitt beschrieben. Das einzige, was
zu diesem Zeitpunkt funktionieren muss, ist, dass Sie die ๐ค Transformers-Implementierung von *brand_new_bert* instanziieren kรถnnen, *d.h.* der
folgende Befehl sollte funktionieren:
```python
from transformers import BrandNewBertModel, BrandNewBertConfig
model = BrandNewBertModel(BrandNewBertConfig())
```
Der obige Befehl erstellt ein Modell gemรคร den Standardparametern, die in `BrandNewBertConfig()` definiert sind, mit
zufรคlligen Gewichten und stellt damit sicher, dass die `init()` Methoden aller Komponenten funktionieren.
Beachten Sie, dass alle zufรคlligen Initialisierungen in der Methode `_init_weights` Ihres `BrandnewBertPreTrainedModel` stattfinden sollten.
Klasse erfolgen sollte. Sie sollte alle Blattmodule in Abhรคngigkeit von den Variablen der Konfiguration initialisieren. Hier ist ein Beispiel mit der
BERT `_init_weights` Methode:
```py
def _init_weights(self, module):
"""Initialize the weights"""
if isinstance(module, nn.Linear):
module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
if module.bias is not None:
module.bias.data.zero_()
elif isinstance(module, nn.Embedding):
module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
if module.padding_idx is not None:
module.weight.data[module.padding_idx].zero_()
elif isinstance(module, nn.LayerNorm):
module.bias.data.zero_()
module.weight.data.fill_(1.0)
```
Sie kรถnnen weitere benutzerdefinierte Schemata verwenden, wenn Sie eine spezielle Initialisierung fรผr einige Module benรถtigen. Zum Beispiel in
`Wav2Vec2ForPreTraining` mรผssen die letzten beiden linearen Schichten die Initialisierung des regulรคren PyTorch `nn.Linear` haben.
aber alle anderen sollten eine Initialisierung wie oben verwenden. Dies ist wie folgt kodiert:
```py
def _init_weights(self, module):
"""Initialize the weights"""
if isinstnace(module, Wav2Vec2ForPreTraining):
module.project_hid.reset_parameters()
module.project_q.reset_parameters()
module.project_hid._is_hf_initialized = True
module.project_q._is_hf_initialized = True
elif isinstance(module, nn.Linear):
module.weight.data.normal_(mean=0.0, std=self.config.initializer_range)
if module.bias is not None:
module.bias.data.zero_()
```
Das Flag `_is_hf_initialized` wird intern verwendet, um sicherzustellen, dass wir ein Submodul nur einmal initialisieren. Wenn Sie es auf
True` fรผr `module.project_q` und `module.project_hid` setzen, stellen wir sicher, dass die benutzerdefinierte Initialisierung, die wir vorgenommen haben, spรคter nicht รผberschrieben wird,
die Funktion `_init_weights` nicht auf sie angewendet wird.
**6. Schreiben Sie ein Konvertierungsskript**
Als nรคchstes sollten Sie ein Konvertierungsskript schreiben, mit dem Sie den Checkpoint, den Sie zum Debuggen von *brand_new_bert* im
im ursprรผnglichen Repository in einen Prรผfpunkt konvertieren, der mit Ihrer gerade erstellten ๐ค Transformers-Implementierung von
*brand_new_bert*. Es ist nicht ratsam, das Konvertierungsskript von Grund auf neu zu schreiben, sondern die bereits
bestehenden Konvertierungsskripten in ๐ค Transformers nach einem Skript zu suchen, das fรผr die Konvertierung eines รคhnlichen Modells verwendet wurde, das im
demselben Framework wie *brand_new_bert* geschrieben wurde. Normalerweise reicht es aus, ein bereits vorhandenes Konvertierungsskript zu kopieren und
es fรผr Ihren Anwendungsfall leicht anzupassen. Zรถgern Sie nicht, das Hugging Face Team zu bitten, Sie auf ein รคhnliches, bereits vorhandenes
Konvertierungsskript fรผr Ihr Modell zu finden.
- Wenn Sie ein Modell von TensorFlow nach PyTorch portieren, ist ein guter Ausgangspunkt das Konvertierungsskript von BERT [hier] (https://github.com/huggingface/transformers/blob/7acfa95afb8194f8f9c1f4d2c6028224dbed35a2/src/transformers/models/bert/modeling_bert.py#L91)
- Wenn Sie ein Modell von PyTorch nach PyTorch portieren, ist ein guter Ausgangspunkt das Konvertierungsskript von BART [hier](https://github.com/huggingface/transformers/blob/main/src/transformers/models/bart/convert_bart_original_pytorch_checkpoint_to_pytorch.py)
Im Folgenden werden wir kurz erklรคren, wie PyTorch-Modelle Ebenengewichte speichern und Ebenennamen definieren. In PyTorch wird der
Name einer Ebene durch den Namen des Klassenattributs definiert, das Sie der Ebene geben. Lassen Sie uns ein Dummy-Modell in
PyTorch, das wir `SimpleModel` nennen, wie folgt:
```python
from torch import nn
class SimpleModel(nn.Module):
def __init__(self):
super().__init__()
self.dense = nn.Linear(10, 10)
self.intermediate = nn.Linear(10, 10)
self.layer_norm = nn.LayerNorm(10)
```
Jetzt kรถnnen wir eine Instanz dieser Modelldefinition erstellen, die alle Gewichte ausfรผllt: `dense`, `intermediate`,
`layer_norm` mit zufรคlligen Gewichten. Wir kรถnnen das Modell ausdrucken, um seine Architektur zu sehen
```python
model = SimpleModel()
print(model)
```
Dies gibt folgendes aus:
```
SimpleModel(
(dense): Linear(in_features=10, out_features=10, bias=True)
(intermediate): Linear(in_features=10, out_features=10, bias=True)
(layer_norm): LayerNorm((10,), eps=1e-05, elementwise_affine=True)
)
```
Wir kรถnnen sehen, dass die Ebenennamen durch den Namen des Klassenattributs in PyTorch definiert sind. Sie kรถnnen die Gewichtswerte
Werte einer bestimmten Ebene anzeigen lassen:
```python
print(model.dense.weight.data)
```
um zu sehen, dass die Gewichte zufรคllig initialisiert wurden
```
tensor([[-0.0818, 0.2207, -0.0749, -0.0030, 0.0045, -0.1569, -0.1598, 0.0212,
-0.2077, 0.2157],
[ 0.1044, 0.0201, 0.0990, 0.2482, 0.3116, 0.2509, 0.2866, -0.2190,
0.2166, -0.0212],
[-0.2000, 0.1107, -0.1999, -0.3119, 0.1559, 0.0993, 0.1776, -0.1950,
-0.1023, -0.0447],
[-0.0888, -0.1092, 0.2281, 0.0336, 0.1817, -0.0115, 0.2096, 0.1415,
-0.1876, -0.2467],
[ 0.2208, -0.2352, -0.1426, -0.2636, -0.2889, -0.2061, -0.2849, -0.0465,
0.2577, 0.0402],
[ 0.1502, 0.2465, 0.2566, 0.0693, 0.2352, -0.0530, 0.1859, -0.0604,
0.2132, 0.1680],
[ 0.1733, -0.2407, -0.1721, 0.1484, 0.0358, -0.0633, -0.0721, -0.0090,
0.2707, -0.2509],
[-0.1173, 0.1561, 0.2945, 0.0595, -0.1996, 0.2988, -0.0802, 0.0407,
0.1829, -0.1568],
[-0.1164, -0.2228, -0.0403, 0.0428, 0.1339, 0.0047, 0.1967, 0.2923,
0.0333, -0.0536],
[-0.1492, -0.1616, 0.1057, 0.1950, -0.2807, -0.2710, -0.1586, 0.0739,
0.2220, 0.2358]]).
```
Im Konvertierungsskript sollten Sie diese zufรคllig initialisierten Gewichte mit den genauen Gewichten der
entsprechenden Ebene im Kontrollpunkt. *Z.B.*
```python
# retrieve matching layer weights, e.g. by
# recursive algorithm
layer_name = "dense"
pretrained_weight = array_of_dense_layer
model_pointer = getattr(model, "dense")
model_pointer.weight.data = torch.from_numpy(pretrained_weight)
```
Dabei mรผssen Sie sicherstellen, dass jedes zufรคllig initialisierte Gewicht Ihres PyTorch-Modells und sein entsprechendes
Checkpoint-Gewicht in **Form und Name** genau รผbereinstimmen. Zu diesem Zweck ist es **notwendig**, assert
Anweisungen fรผr die Form hinzuzufรผgen und die Namen der Checkpoint-Gewichte auszugeben. Sie sollten z.B. Anweisungen hinzufรผgen wie:
```python
assert (
model_pointer.weight.shape == pretrained_weight.shape
), f"Pointer shape of random weight {model_pointer.shape} and array shape of checkpoint weight {pretrained_weight.shape} mismatched"
```
Auรerdem sollten Sie die Namen der beiden Gewichte ausdrucken, um sicherzustellen, dass sie รผbereinstimmen, *z.B.*.
```python
logger.info(f"Initialize PyTorch weight {layer_name} from {pretrained_weight.name}")
```
Wenn entweder die Form oder der Name nicht รผbereinstimmt, haben Sie wahrscheinlich das falsche Kontrollpunktgewicht einer zufรคllig
Ebene der ๐ค Transformers-Implementierung zugewiesen.
Eine falsche Form ist hรถchstwahrscheinlich auf eine falsche Einstellung der Konfigurationsparameter in `BrandNewBertConfig()` zurรผckzufรผhren, die
nicht genau mit denen รผbereinstimmen, die fรผr den zu konvertierenden Prรผfpunkt verwendet wurden. Es kรถnnte aber auch sein, dass
die PyTorch-Implementierung eines Layers erfordert, dass das Gewicht vorher transponiert wird.
Schlieรlich sollten Sie auch รผberprรผfen, ob **alle** erforderlichen Gewichte initialisiert sind und alle Checkpoint-Gewichte ausgeben, die
die nicht zur Initialisierung verwendet wurden, um sicherzustellen, dass das Modell korrekt konvertiert wurde. Es ist vรถllig normal, dass die
Konvertierungsversuche entweder mit einer falschen Shape-Anweisung oder einer falschen Namenszuweisung fehlschlagen. Das liegt hรถchstwahrscheinlich daran, dass entweder
Sie haben falsche Parameter in `BrandNewBertConfig()` verwendet, haben eine falsche Architektur in der ๐ค Transformers
Implementierung, Sie haben einen Fehler in den `init()` Funktionen einer der Komponenten der ๐ค Transformers
Implementierung oder Sie mรผssen eine der Kontrollpunktgewichte transponieren.
Dieser Schritt sollte mit dem vorherigen Schritt wiederholt werden, bis alle Gewichte des Kontrollpunkts korrekt in das
Transformers-Modell geladen sind. Nachdem Sie den Prรผfpunkt korrekt in die ๐ค Transformers-Implementierung geladen haben, kรถnnen Sie das Modell
das Modell unter einem Ordner Ihrer Wahl `/path/to/converted/checkpoint/folder` speichern, der dann sowohl ein
Datei `pytorch_model.bin` und eine Datei `config.json` enthalten sollte:
```python
model.save_pretrained("/path/to/converted/checkpoint/folder")
```
**7. Implementieren Sie den Vorwรคrtspass**
Nachdem es Ihnen gelungen ist, die trainierten Gewichte korrekt in die ๐ค Transformers-Implementierung zu laden, sollten Sie nun dafรผr sorgen
sicherstellen, dass der Forward Pass korrekt implementiert ist. In [Machen Sie sich mit dem ursprรผnglichen Repository vertraut](#34-run-a-pretrained-checkpoint-using-the-original-repository) haben Sie bereits ein Skript erstellt, das einen Forward Pass
Durchlauf des Modells unter Verwendung des Original-Repositorys durchfรผhrt. Jetzt sollten Sie ein analoges Skript schreiben, das die ๐ค Transformers
Implementierung anstelle der Originalimplementierung verwenden. Es sollte wie folgt aussehen:
```python
model = BrandNewBertModel.from_pretrained("/path/to/converted/checkpoint/folder")
input_ids = [0, 4, 4, 3, 2, 4, 1, 7, 19]
output = model(input_ids).last_hidden_states
```
Es ist sehr wahrscheinlich, dass die ๐ค Transformers-Implementierung und die ursprรผngliche Modell-Implementierung nicht genau die gleiche Ausgabe liefern.
beim ersten Mal nicht die gleiche Ausgabe liefern oder dass der Vorwรคrtsdurchlauf einen Fehler auslรถst. Seien Sie nicht enttรคuscht - das ist zu erwarten! Erstens,
sollten Sie sicherstellen, dass der Vorwรคrtsdurchlauf keine Fehler auslรถst. Es passiert oft, dass die falschen Dimensionen verwendet werden
verwendet werden, was zu einem *Dimensionality mismatch* Fehler fรผhrt oder dass der falsche Datentyp verwendet wird, *z.B.* `torch.long`
anstelle von `torch.float32`. Zรถgern Sie nicht, das Hugging Face Team um Hilfe zu bitten, wenn Sie bestimmte Fehler nicht lรถsen kรถnnen.
bestimmte Fehler nicht lรถsen kรถnnen.
Um sicherzustellen, dass die Implementierung von ๐ค Transformers korrekt funktioniert, mรผssen Sie sicherstellen, dass die Ausgaben
einer Genauigkeit von `1e-3` entsprechen. Zunรคchst sollten Sie sicherstellen, dass die Ausgabeformen identisch sind, *d.h.*.
Die Ausgabeform *outputs.shape* sollte fรผr das Skript der ๐ค Transformers-Implementierung und die ursprรผngliche
Implementierung ergeben. Als nรคchstes sollten Sie sicherstellen, dass auch die Ausgabewerte identisch sind. Dies ist einer der schwierigsten
Teile des Hinzufรผgens eines neuen Modells. Hรคufige Fehler, warum die Ausgaben nicht identisch sind, sind:
- Einige Ebenen wurden nicht hinzugefรผgt, *d.h.* eine *Aktivierungsebene* wurde nicht hinzugefรผgt, oder die Restverbindung wurde vergessen
- Die Worteinbettungsmatrix wurde nicht gebunden
- Es werden die falschen Positionseinbettungen verwendet, da die ursprรผngliche Implementierung einen Offset verwendet
- Dropout wird wรคhrend des Vorwรคrtsdurchlaufs angewendet. Um dies zu beheben, stellen Sie sicher, dass *model.training auf False* steht und dass keine Dropout
Schicht wรคhrend des Vorwรคrtsdurchlaufs fรคlschlicherweise aktiviert wird, *d.h.* รผbergeben Sie *self.training* an [PyTorch's functional dropout](https://pytorch.org/docs/stable/nn.functional.html?highlight=dropout#torch.nn.functional.dropout)
Der beste Weg, das Problem zu beheben, besteht normalerweise darin, sich den Vorwรคrtsdurchlauf der ursprรผnglichen Implementierung und die ๐ค
Transformers-Implementierung nebeneinander zu sehen und zu prรผfen, ob es Unterschiede gibt. Idealerweise sollten Sie die
Zwischenergebnisse beider Implementierungen des Vorwรคrtsdurchlaufs debuggen/ausdrucken, um die genaue Position im Netzwerk zu finden, an der die ๐ค
Transformers-Implementierung eine andere Ausgabe zeigt als die ursprรผngliche Implementierung. Stellen Sie zunรคchst sicher, dass die
hartcodierten `input_ids` in beiden Skripten identisch sind. รberprรผfen Sie dann, ob die Ausgaben der ersten Transformation von
der `input_ids` (normalerweise die Worteinbettungen) identisch sind. Und dann arbeiten Sie sich bis zur allerletzten Schicht des
Netzwerks. Irgendwann werden Sie einen Unterschied zwischen den beiden Implementierungen feststellen, der Sie auf den Fehler
in der Implementierung von ๐ค Transformers hinweist. Unserer Erfahrung nach ist ein einfacher und effizienter Weg, viele Druckanweisungen hinzuzufรผgen
sowohl in der Original-Implementierung als auch in der ๐ค Transformers-Implementierung an den gleichen Stellen im Netzwerk
hinzuzufรผgen und nacheinander Druckanweisungen zu entfernen, die dieselben Werte fรผr Zwischenprรคsentationen anzeigen.
Wenn Sie sicher sind, dass beide Implementierungen die gleiche Ausgabe liefern, รผberprรผfen Sie die Ausgaben mit
`torch.allclose(original_output, output, atol=1e-3)` รผberprรผfen, haben Sie den schwierigsten Teil hinter sich! Herzlichen Glรผckwunsch - die
Arbeit, die noch zu erledigen ist, sollte ein Kinderspiel sein ๐.
**8. Hinzufรผgen aller notwendigen Modelltests**
An diesem Punkt haben Sie erfolgreich ein neues Modell hinzugefรผgt. Es ist jedoch sehr gut mรถglich, dass das Modell noch nicht
noch nicht vollstรคndig mit dem erforderlichen Design รผbereinstimmt. Um sicherzustellen, dass die Implementierung vollstรคndig kompatibel mit ๐ค Transformers ist, sollten alle
gemeinsamen Tests bestehen. Der Cookiecutter sollte automatisch eine Testdatei fรผr Ihr Modell hinzugefรผgt haben, wahrscheinlich unter
demselben `tests/models/brand_new_bert/test_modeling_brand_new_bert.py`. Fรผhren Sie diese Testdatei aus, um zu รผberprรผfen, ob alle gรคngigen
Tests bestehen:
```bash
pytest tests/models/brand_new_bert/test_modeling_brand_new_bert.py
```
Nachdem Sie alle allgemeinen Tests festgelegt haben, mรผssen Sie nun sicherstellen, dass all die schรถne Arbeit, die Sie geleistet haben, gut getestet ist, damit
- a) die Community Ihre Arbeit leicht nachvollziehen kann, indem sie sich spezifische Tests von *brand_new_bert* ansieht
- b) zukรผnftige รnderungen an Ihrem Modell keine wichtigen Funktionen des Modells zerstรถren.
Als erstes sollten Sie Integrationstests hinzufรผgen. Diese Integrationstests tun im Wesentlichen dasselbe wie die Debugging-Skripte
die Sie zuvor zur Implementierung des Modells in ๐ค Transformers verwendet haben. Eine Vorlage fรผr diese Modelltests wurde bereits von dem
Cookiecutter hinzugefรผgt, die `BrandNewBertModelIntegrationTests` heiรt und nur noch von Ihnen ausgefรผllt werden muss. Um sicherzustellen, dass diese
Tests erfolgreich sind, fรผhren Sie
```bash
RUN_SLOW=1 pytest -sv tests/models/brand_new_bert/test_modeling_brand_new_bert.py::BrandNewBertModelIntegrationTests
```
<Tip>
Falls Sie Windows verwenden, sollten Sie `RUN_SLOW=1` durch `SET RUN_SLOW=1` ersetzen.
</Tip>
Zweitens sollten alle Funktionen, die speziell fรผr *brand_new_bert* sind, zusรคtzlich in einem separaten Test getestet werden unter
`BrandNewBertModelTester`/``BrandNewBertModelTest`. Dieser Teil wird oft vergessen, ist aber in zweierlei Hinsicht รคuรerst nรผtzlich
Weise:
- Er hilft dabei, das Wissen, das Sie wรคhrend der Modellerweiterung erworben haben, an die Community weiterzugeben, indem er zeigt, wie die
speziellen Funktionen von *brand_new_bert* funktionieren sollten.
- Kรผnftige Mitwirkende kรถnnen รnderungen am Modell schnell testen, indem sie diese speziellen Tests ausfรผhren.
**9. Implementieren Sie den Tokenizer**
Als nรคchstes sollten wir den Tokenizer von *brand_new_bert* hinzufรผgen. Normalerweise ist der Tokenizer รคquivalent oder sehr รคhnlich zu einem
bereits vorhandenen Tokenizer von ๐ค Transformers.
Es ist sehr wichtig, die ursprรผngliche Tokenizer-Datei zu finden/extrahieren und es zu schaffen, diese Datei in die ๐ค
Transformers Implementierung des Tokenizers zu laden.
Um sicherzustellen, dass der Tokenizer korrekt funktioniert, empfiehlt es sich, zunรคchst ein Skript im ursprรผnglichen Repository zu erstellen
zu erstellen, das eine Zeichenkette eingibt und die `input_ids` zurรผckgibt. Es kรถnnte etwa so aussehen (in Pseudocode):
```python
input_str = "This is a long example input string containing special characters .$?-, numbers 2872 234 12 and words."
model = BrandNewBertModel.load_pretrained_checkpoint("/path/to/checkpoint/")
input_ids = model.tokenize(input_str)
```
Mรถglicherweise mรผssen Sie noch einmal einen Blick in das ursprรผngliche Repository werfen, um die richtige Tokenizer-Funktion zu finden, oder Sie mรผssen
Sie mรผssen vielleicht sogar รnderungen an Ihrem Klon des Original-Repositorys vornehmen, um nur die `input_ids` auszugeben. Nach dem Schreiben
ein funktionierendes Tokenisierungsskript geschrieben, das das ursprรผngliche Repository verwendet, sollten Sie ein analoges Skript fรผr ๐ค Transformers
erstellt werden. Es sollte รคhnlich wie dieses aussehen:
```python
from transformers import BrandNewBertTokenizer
input_str = "This is a long example input string containing special characters .$?-, numbers 2872 234 12 and words."
tokenizer = BrandNewBertTokenizer.from_pretrained("/path/to/tokenizer/folder/")
input_ids = tokenizer(input_str).input_ids
```
Wenn beide `input_ids` die gleichen Werte ergeben, sollte als letzter Schritt auch eine Tokenizer-Testdatei hinzugefรผgt werden.
Analog zu den Modellierungstestdateien von *brand_new_bert* sollten auch die Tokenisierungs-Testdateien von *brand_new_bert*
eine Reihe von fest kodierten Integrationstests enthalten.
**10. Fรผhren Sie End-to-End-Integrationstests aus**
Nachdem Sie den Tokenizer hinzugefรผgt haben, sollten Sie auch ein paar End-to-End-Integrationstests, die sowohl das Modell als auch den
Tokenizer zu `tests/models/brand_new_bert/test_modeling_brand_new_bert.py` in ๐ค Transformers.
Ein solcher Test sollte bei einem aussagekrรคftigen
Text-zu-Text-Beispiel zeigen, dass die Implementierung von ๐ค Transformers wie erwartet funktioniert. Ein aussagekrรคftiges Text-zu-Text-Beispiel kann
z.B. *ein Quell-zu-Ziel-รbersetzungspaar, ein Artikel-zu-Zusammenfassung-Paar, ein Frage-zu-Antwort-Paar, usw... Wenn keiner der
der portierten Prรผfpunkte in einer nachgelagerten Aufgabe feinabgestimmt wurde, genรผgt es, sich einfach auf die Modelltests zu verlassen. In einem
letzten Schritt, um sicherzustellen, dass das Modell voll funktionsfรคhig ist, sollten Sie alle Tests auch auf der GPU durchfรผhren. Es kann
Es kann vorkommen, dass Sie vergessen haben, einige `.to(self.device)` Anweisungen zu internen Tensoren des Modells hinzuzufรผgen, was in einem solchen
Test zu einem Fehler fรผhren wรผrde. Falls Sie keinen Zugang zu einem Grafikprozessor haben, kann das Hugging Face Team diese Tests fรผr Sie durchfรผhren.
Tests fรผr Sie รผbernehmen.
**11. Docstring hinzufรผgen**
Nun sind alle notwendigen Funktionen fรผr *brand_new_bert* hinzugefรผgt - Sie sind fast fertig! Das Einzige, was Sie noch hinzufรผgen mรผssen, ist
ein schรถner Docstring und eine Doku-Seite. Der Cookiecutter sollte eine Vorlagendatei namens
`docs/source/model_doc/brand_new_bert.md` hinzugefรผgt haben, die Sie ausfรผllen sollten. Die Benutzer Ihres Modells werden in der Regel zuerst einen Blick auf
diese Seite ansehen, bevor sie Ihr Modell verwenden. Daher muss die Dokumentation verstรคndlich und prรคgnant sein. Es ist sehr nรผtzlich fรผr
die Gemeinschaft, einige *Tipps* hinzuzufรผgen, um zu zeigen, wie das Modell verwendet werden sollte. Zรถgern Sie nicht, das Hugging Face-Team anzupingen
bezรผglich der Docstrings.
Stellen Sie als nรคchstes sicher, dass der zu `src/transformers/models/brand_new_bert/modeling_brand_new_bert.py` hinzugefรผgte docstring
korrekt ist und alle erforderlichen Eingaben und Ausgaben enthรคlt. Wir haben eine ausfรผhrliche Anleitung zum Schreiben von Dokumentationen und unserem Docstring-Format [hier](writing-documentation). Es ist immer gut, sich daran zu erinnern, dass die Dokumentation
mindestens so sorgfรคltig behandelt werden sollte wie der Code in ๐ค Transformers, denn die Dokumentation ist in der Regel der erste Kontaktpunkt der
Berรผhrungspunkt der Community mit dem Modell ist.
**Code refactor**
Groรartig, jetzt haben Sie den gesamten erforderlichen Code fรผr *brand_new_bert* hinzugefรผgt. An diesem Punkt sollten Sie einige mรถgliche
falschen Codestil korrigieren, indem Sie ausfรผhren:
```bash
make style
```
und รผberprรผfen Sie, ob Ihr Kodierungsstil die Qualitรคtsprรผfung besteht:
```bash
make quality
```
Es gibt noch ein paar andere sehr strenge Designtests in ๐ค Transformers, die mรถglicherweise noch fehlschlagen, was sich in den
den Tests Ihres Pull Requests. Dies liegt oft an fehlenden Informationen im Docstring oder an einer falschen
Benennung. Das Hugging Face Team wird Ihnen sicherlich helfen, wenn Sie hier nicht weiterkommen.
Und schlieรlich ist es immer eine gute Idee, den eigenen Code zu refaktorisieren, nachdem man sichergestellt hat, dass er korrekt funktioniert. Wenn alle
Tests bestanden haben, ist es nun an der Zeit, den hinzugefรผgten Code noch einmal durchzugehen und einige รberarbeitungen vorzunehmen.
Sie haben nun den Codierungsteil abgeschlossen, herzlichen Glรผckwunsch! ๐ Sie sind groรartig! ๐
**12. Laden Sie die Modelle in den Model Hub hoch**
In diesem letzten Teil sollten Sie alle Checkpoints konvertieren und in den Modell-Hub hochladen und eine Modellkarte fรผr jeden
hochgeladenen Modell-Kontrollpunkt. Sie kรถnnen sich mit den Hub-Funktionen vertraut machen, indem Sie unsere [Model sharing and uploading Page](model_sharing) lesen. Hier sollten Sie mit dem Hugging Face-Team zusammenarbeiten, um einen passenden Namen fรผr jeden
Checkpoint festzulegen und die erforderlichen Zugriffsrechte zu erhalten, um das Modell unter der Organisation des Autors *brand_new_bert* hochladen zu kรถnnen.
*brand_new_bert*. Die Methode `push_to_hub`, die in allen Modellen in `transformers` vorhanden ist, ist ein schneller und effizienter Weg, Ihren Checkpoint in den Hub zu pushen. Ein kleines Snippet ist unten eingefรผgt:
```python
brand_new_bert.push_to_hub("brand_new_bert")
# Uncomment the following line to push to an organization.
# brand_new_bert.push_to_hub("<organization>/brand_new_bert")
```
Es lohnt sich, etwas Zeit darauf zu verwenden, fรผr jeden Kontrollpunkt passende Musterkarten zu erstellen. Die Modellkarten sollten die
spezifischen Merkmale dieses bestimmten Prรผfpunkts hervorheben, * z.B.* auf welchem Datensatz wurde der Prรผfpunkt
vortrainiert/abgestimmt? Fรผr welche nachgelagerte Aufgabe sollte das Modell verwendet werden? Und fรผgen Sie auch etwas Code bei, wie Sie
wie das Modell korrekt verwendet wird.
**13. (Optional) Notizbuch hinzufรผgen**
Es ist sehr hilfreich, ein Notizbuch hinzuzufรผgen, in dem im Detail gezeigt wird, wie *brand_new_bert* fรผr Schlussfolgerungen verwendet werden kann und/oder
bei einer nachgelagerten Aufgabe feinabgestimmt wird. Dies ist nicht zwingend erforderlich, um Ihren PR zusammenzufรผhren, aber sehr nรผtzlich fรผr die Gemeinschaft.
**14. Reichen Sie Ihren fertigen PR ein**
Sie sind jetzt mit der Programmierung fertig und kรถnnen zum letzten Schritt รผbergehen, nรคmlich der Zusammenfรผhrung Ihres PR mit main. Normalerweise hat das
Hugging Face Team Ihnen an diesem Punkt bereits geholfen haben, aber es lohnt sich, sich etwas Zeit zu nehmen, um Ihrem fertigen
PR eine schรถne Beschreibung zu geben und eventuell Kommentare zu Ihrem Code hinzuzufรผgen, wenn Sie Ihren Gutachter auf bestimmte Designentscheidungen hinweisen wollen.
Gutachter hinweisen wollen.
### Teilen Sie Ihre Arbeit!!
Jetzt ist es an der Zeit, von der Community Anerkennung fรผr Ihre Arbeit zu bekommen! Die Fertigstellung einer Modellergรคnzung ist ein wichtiger
Beitrag zu Transformers und der gesamten NLP-Gemeinschaft. Ihr Code und die portierten vortrainierten Modelle werden sicherlich
von Hunderten und vielleicht sogar Tausenden von Entwicklern und Forschern genutzt werden. Sie sollten stolz auf Ihre Arbeit sein und Ihre
Ihre Leistung mit der Gemeinschaft teilen.
**Sie haben ein weiteres Modell erstellt, das fรผr jeden in der Community super einfach zugรคnglich ist! ๐คฏ**
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/run_scripts.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Trainieren mit einem Skript
Neben den ๐ค Transformers [notebooks](./noteboks/README) gibt es auch Beispielskripte, die zeigen, wie man ein Modell fรผr eine Aufgabe mit [PyTorch](https://github.com/huggingface/transformers/tree/main/examples/pytorch), [TensorFlow](https://github.com/huggingface/transformers/tree/main/examples/tensorflow) oder [JAX/Flax](https://github.com/huggingface/transformers/tree/main/examples/flax) trainiert.
Sie werden auch Skripte finden, die wir in unseren [Forschungsprojekten](https://github.com/huggingface/transformers/tree/main/examples/research_projects) und [Legacy-Beispielen](https://github.com/huggingface/transformers/tree/main/examples/legacy) verwendet haben und die grรถรtenteils von der Community stammen. Diese Skripte werden nicht aktiv gepflegt und erfordern eine bestimmte Version von ๐ค Transformers, die hรถchstwahrscheinlich nicht mit der neuesten Version der Bibliothek kompatibel ist.
Es wird nicht erwartet, dass die Beispielskripte bei jedem Problem sofort funktionieren. Mรถglicherweise mรผssen Sie das Skript an das Problem anpassen, das Sie zu lรถsen versuchen. Um Ihnen dabei zu helfen, legen die meisten Skripte vollstรคndig offen, wie die Daten vorverarbeitet werden, so dass Sie sie nach Bedarf fรผr Ihren Anwendungsfall bearbeiten kรถnnen.
Fรผr jede Funktion, die Sie in einem Beispielskript implementieren mรถchten, diskutieren Sie bitte im [Forum] (https://discuss.huggingface.co/) oder in einem [issue] (https://github.com/huggingface/transformers/issues), bevor Sie einen Pull Request einreichen. Wir freuen uns zwar รผber Fehlerkorrekturen, aber es ist unwahrscheinlich, dass wir einen Pull Request zusammenfรผhren, der mehr Funktionalitรคt auf Kosten der Lesbarkeit hinzufรผgt.
Diese Anleitung zeigt Ihnen, wie Sie ein Beispiel fรผr ein Trainingsskript zur Zusammenfassung in [PyTorch](https://github.com/huggingface/transformers/tree/main/examples/pytorch/summarization) und [TensorFlow](https://github.com/huggingface/transformers/tree/main/examples/tensorflow/summarization) ausfรผhren kรถnnen. Sofern nicht anders angegeben, sollten alle Beispiele mit beiden Frameworks funktionieren.
## Einrichtung
Um die neueste Version der Beispielskripte erfolgreich auszufรผhren, **mรผssen Sie ๐ค Transformers aus dem Quellcode** in einer neuen virtuellen Umgebung installieren:
```bash
git clone https://github.com/huggingface/transformers
cd transformers
pip install .
```
Fรผr รคltere Versionen der Beispielskripte klicken Sie auf die Umschalttaste unten:
<details>
<summary>Beispiele fรผr รคltere Versionen von ๐ค Transformers</summary>
<ul>
<li><a href="https://github.com/huggingface/transformers/tree/v4.5.1/examples">v4.5.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v4.4.2/examples">v4.4.2</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v4.3.3/examples">v4.3.3</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v4.2.2/examples">v4.2.2</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v4.1.1/examples">v4.1.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v4.0.1/examples">v4.0.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.5.1/examples">v3.5.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.4.0/examples">v3.4.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.3.1/examples">v3.3.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.2.0/examples">v3.2.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.1.0/examples">v3.1.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.0.2/examples">v3.0.2</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.11.0/examples">v2.11.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.10.0/examples">v2.10.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.9.1/examples">v2.9.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.8.0/examples">v2.8.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.7.0/examples">v2.7.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.6.0/examples">v2.6.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.5.1/examples">v2.5.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.4.0/examples">v2.4.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.3.0/examples">v2.3.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.2.0/examples">v2.2.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.1.0/examples">v2.1.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.0.0/examples">v2.0.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v1.2.0/examples">v1.2.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v1.1.0/examples">v1.1.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v1.0.0/examples">v1.0.0</a></li>
</ul>
</details>
Dann stellen Sie Ihren aktuellen Klon von ๐ค Transformers auf eine bestimmte Version um, z.B. v3.5.1:
```bash
git checkout tags/v3.5.1
```
Nachdem Sie die richtige Bibliotheksversion eingerichtet haben, navigieren Sie zu dem Beispielordner Ihrer Wahl und installieren die beispielspezifischen Anforderungen:
```bash
pip install -r requirements.txt
```
## Ein Skript ausfรผhren
<frameworkcontent>
<pt>
Das Beispielskript lรคdt einen Datensatz aus der ๐ค [Datasets](https://huggingface.co/docs/datasets/) Bibliothek herunter und verarbeitet ihn vor. Dann nimmt das Skript eine Feinabstimmung eines Datensatzes mit dem [Trainer](https://huggingface.co/docs/transformers/main_classes/trainer) auf einer Architektur vor, die eine Zusammenfassung unterstรผtzt. Das folgende Beispiel zeigt, wie die Feinabstimmung von [T5-small](https://huggingface.co/t5-small) auf dem Datensatz [CNN/DailyMail](https://huggingface.co/datasets/cnn_dailymail) durchgefรผhrt wird. Das T5-Modell benรถtigt aufgrund der Art und Weise, wie es trainiert wurde, ein zusรคtzliches Argument `source_prefix`. Mit dieser Eingabeaufforderung weiร T5, dass es sich um eine Zusammenfassungsaufgabe handelt.
```bash
python examples/pytorch/summarization/run_summarization.py \
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--predict_with_generate
```
</pt>
<tf>
Das Beispielskript lรคdt einen Datensatz aus der ๐ค [Datasets](https://huggingface.co/docs/datasets/) Bibliothek herunter und verarbeitet ihn vor. Anschlieรend nimmt das Skript die Feinabstimmung eines Datensatzes mit Keras auf einer Architektur vor, die die Zusammenfassung unterstรผtzt. Das folgende Beispiel zeigt, wie die Feinabstimmung von [T5-small](https://huggingface.co/t5-small) auf dem [CNN/DailyMail](https://huggingface.co/datasets/cnn_dailymail) Datensatz durchgefรผhrt wird. Das T5-Modell benรถtigt aufgrund der Art und Weise, wie es trainiert wurde, ein zusรคtzliches Argument `source_prefix`. Mit dieser Eingabeaufforderung weiร T5, dass es sich um eine Zusammenfassungsaufgabe handelt.
```bash
python examples/tensorflow/summarization/run_summarization.py \
--model_name_or_path t5-small \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size 8 \
--per_device_eval_batch_size 16 \
--num_train_epochs 3 \
--do_train \
--do_eval
```
</tf>
</frameworkcontent>
## Verteiltes Training und gemischte Prรคzision
Der [Trainer](https://huggingface.co/docs/transformers/main_classes/trainer) unterstรผtzt verteiltes Training und gemischte Prรคzision, d.h. Sie kรถnnen ihn auch in einem Skript verwenden. So aktivieren Sie diese beiden Funktionen:
- Fรผgen Sie das Argument `fp16` hinzu, um gemischte Genauigkeit zu aktivieren.
- Legen Sie die Anzahl der zu verwendenden GPUs mit dem Argument `nproc_per_node` fest.
```bash
torchrun \
--nproc_per_node 8 pytorch/summarization/run_summarization.py \
--fp16 \
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--predict_with_generate
```
TensorFlow-Skripte verwenden eine [`MirroredStrategy`](https://www.tensorflow.org/guide/distributed_training#mirroredstrategy) fรผr verteiltes Training, und Sie mรผssen dem Trainingsskript keine zusรคtzlichen Argumente hinzufรผgen. Das TensorFlow-Skript verwendet standardmรครig mehrere GPUs, wenn diese verfรผgbar sind.
## Ein Skript auf einer TPU ausfรผhren
<frameworkcontent>
<pt>
Tensor Processing Units (TPUs) sind speziell fรผr die Beschleunigung der Leistung konzipiert. PyTorch unterstรผtzt TPUs mit dem [XLA](https://www.tensorflow.org/xla) Deep Learning Compiler (siehe [hier](https://github.com/pytorch/xla/blob/master/README.md) fรผr weitere Details). Um eine TPU zu verwenden, starten Sie das Skript `xla_spawn.py` und verwenden das Argument `num_cores`, um die Anzahl der TPU-Kerne festzulegen, die Sie verwenden mรถchten.
```bash
python xla_spawn.py --num_cores 8 \
summarization/run_summarization.py \
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--predict_with_generate
```
</pt>
<tf>
Tensor Processing Units (TPUs) sind speziell fรผr die Beschleunigung der Leistung konzipiert. TensorFlow Skripte verwenden eine [`TPUStrategy`](https://www.tensorflow.org/guide/distributed_training#tpustrategy) fรผr das Training auf TPUs. Um eine TPU zu verwenden, รผbergeben Sie den Namen der TPU-Ressource an das Argument `tpu`.
```bash
python run_summarization.py \
--tpu name_of_tpu_resource \
--model_name_or_path t5-small \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size 8 \
--per_device_eval_batch_size 16 \
--num_train_epochs 3 \
--do_train \
--do_eval
```
</tf>
</frameworkcontent>
## Fรผhren Sie ein Skript mit ๐ค Accelerate aus.
๐ค [Accelerate](https://huggingface.co/docs/accelerate) ist eine reine PyTorch-Bibliothek, die eine einheitliche Methode fรผr das Training eines Modells auf verschiedenen Arten von Setups (nur CPU, mehrere GPUs, TPUs) bietet und dabei die vollstรคndige Transparenz der PyTorch-Trainingsschleife beibehรคlt. Stellen Sie sicher, dass Sie ๐ค Accelerate installiert haben, wenn Sie es nicht bereits haben:
> Hinweis: Da Accelerate schnell weiterentwickelt wird, muss die Git-Version von Accelerate installiert sein, um die Skripte auszufรผhren.
```bash
pip install git+https://github.com/huggingface/accelerate
```
Anstelle des Skripts `run_summarization.py` mรผssen Sie das Skript `run_summarization_no_trainer.py` verwenden. Die von Accelerate unterstรผtzten Skripte haben eine Datei `task_no_trainer.py` im Ordner. Beginnen Sie mit dem folgenden Befehl, um eine Konfigurationsdatei zu erstellen und zu speichern:
```bash
accelerate config
```
Testen Sie Ihre Einrichtung, um sicherzustellen, dass sie korrekt konfiguriert ist:
```bash
accelerate test
```
Jetzt sind Sie bereit, das Training zu starten:
```bash
accelerate launch run_summarization_no_trainer.py \
--model_name_or_path t5-small \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir ~/tmp/tst-summarization
```
## Verwenden Sie einen benutzerdefinierten Datensatz
Das Verdichtungsskript unterstรผtzt benutzerdefinierte Datensรคtze, solange es sich um eine CSV- oder JSON-Line-Datei handelt. Wenn Sie Ihren eigenen Datensatz verwenden, mรผssen Sie mehrere zusรคtzliche Argumente angeben:
- `train_file` und `validation_file` geben den Pfad zu Ihren Trainings- und Validierungsdateien an.
- text_column` ist der Eingabetext, der zusammengefasst werden soll.
- Summary_column" ist der auszugebende Zieltext.
Ein Zusammenfassungsskript, das einen benutzerdefinierten Datensatz verwendet, wรผrde wie folgt aussehen:
```bash
python examples/pytorch/summarization/run_summarization.py \
--model_name_or_path t5-small \
--do_train \
--do_eval \
--train_file path_to_csv_or_jsonlines_file \
--validation_file path_to_csv_or_jsonlines_file \
--text_column text_column_name \
--summary_column summary_column_name \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--overwrite_output_dir \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--predict_with_generate
```
## Testen Sie ein Skript
Es ist oft eine gute Idee, Ihr Skript an einer kleineren Anzahl von Beispielen fรผr Datensรคtze auszufรผhren, um sicherzustellen, dass alles wie erwartet funktioniert, bevor Sie sich auf einen ganzen Datensatz festlegen, dessen Fertigstellung Stunden dauern kann. Verwenden Sie die folgenden Argumente, um den Datensatz auf eine maximale Anzahl von Stichproben zu beschrรคnken:
- `max_train_samples`
- `max_eval_samples`
- `max_predict_samples`
```bash
python examples/pytorch/summarization/run_summarization.py \
--model_name_or_path t5-small \
--max_train_samples 50 \
--max_eval_samples 50 \
--max_predict_samples 50 \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--predict_with_generate
```
Nicht alle Beispielskripte unterstรผtzen das Argument `max_predict_samples`. Wenn Sie sich nicht sicher sind, ob Ihr Skript dieses Argument unterstรผtzt, fรผgen Sie das Argument `-h` hinzu, um dies zu รผberprรผfen:
```bash
examples/pytorch/summarization/run_summarization.py -h
```
## Training vom Kontrollpunkt fortsetzen
Eine weitere hilfreiche Option, die Sie aktivieren kรถnnen, ist die Wiederaufnahme des Trainings von einem frรผheren Kontrollpunkt aus. Auf diese Weise kรถnnen Sie im Falle einer Unterbrechung Ihres Trainings dort weitermachen, wo Sie aufgehรถrt haben, ohne von vorne beginnen zu mรผssen. Es gibt zwei Methoden, um das Training von einem Kontrollpunkt aus wieder aufzunehmen.
Die erste Methode verwendet das Argument `output_dir previous_output_dir`, um das Training ab dem letzten in `output_dir` gespeicherten Kontrollpunkt wieder aufzunehmen. In diesem Fall sollten Sie `overwrite_output_dir` entfernen:
```bash
python examples/pytorch/summarization/run_summarization.py
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--output_dir previous_output_dir \
--predict_with_generate
```
Die zweite Methode verwendet das Argument `Resume_from_checkpoint path_to_specific_checkpoint`, um das Training ab einem bestimmten Checkpoint-Ordner wieder aufzunehmen.
```bash
python examples/pytorch/summarization/run_summarization.py
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--resume_from_checkpoint path_to_specific_checkpoint \
--predict_with_generate
```
## Teilen Sie Ihr Modell
Alle Skripte kรถnnen Ihr endgรผltiges Modell in den [Model Hub](https://huggingface.co/models) hochladen. Stellen Sie sicher, dass Sie bei Hugging Face angemeldet sind, bevor Sie beginnen:
```bash
huggingface-cli login
```
Dann fรผgen Sie dem Skript das Argument `push_to_hub` hinzu. Mit diesem Argument wird ein Repository mit Ihrem Hugging Face-Benutzernamen und dem in `output_dir` angegebenen Ordnernamen erstellt.
Wenn Sie Ihrem Repository einen bestimmten Namen geben mรถchten, fรผgen Sie ihn mit dem Argument `push_to_hub_model_id` hinzu. Das Repository wird automatisch unter Ihrem Namensraum aufgefรผhrt.
Das folgende Beispiel zeigt, wie Sie ein Modell mit einem bestimmten Repository-Namen hochladen kรถnnen:
```bash
python examples/pytorch/summarization/run_summarization.py
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--push_to_hub \
--push_to_hub_model_id finetuned-t5-cnn_dailymail \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--predict_with_generate
```
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/_config.py
|
# docstyle-ignore
INSTALL_CONTENT = """
# Transformers installation
! pip install transformers datasets
# To install from source instead of the last release, comment the command above and uncomment the following one.
# ! pip install git+https://github.com/huggingface/transformers.git
"""
notebook_first_cells = [{"type": "code", "content": INSTALL_CONTENT}]
black_avoid_patterns = {
"{processor_class}": "FakeProcessorClass",
"{model_class}": "FakeModelClass",
"{object_class}": "FakeObjectClass",
}
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/autoclass_tutorial.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Vortrainierte Instanzen mit einer AutoClass laden
Bei so vielen verschiedenen Transformator-Architekturen kann es eine Herausforderung sein, eine fรผr Ihren Checkpoint zu erstellen. Als Teil der ๐ค Transformers Kernphilosophie, die Bibliothek leicht, einfach und flexibel nutzbar zu machen, leitet eine `AutoClass` automatisch die richtige Architektur aus einem gegebenen Checkpoint ab und lรคdt sie. Mit der Methode `from_pretrained()` kann man schnell ein vortrainiertes Modell fรผr eine beliebige Architektur laden, so dass man keine Zeit und Ressourcen aufwenden muss, um ein Modell von Grund auf zu trainieren. Die Erstellung dieser Art von Checkpoint-agnostischem Code bedeutet, dass Ihr Code, wenn er fรผr einen Checkpoint funktioniert, auch mit einem anderen Checkpoint funktionieren wird - solange er fรผr eine รคhnliche Aufgabe trainiert wurde - selbst wenn die Architektur unterschiedlich ist.
<Tip>
Denken Sie daran, dass sich die Architektur auf das Skelett des Modells bezieht und die Checkpoints die Gewichte fรผr eine bestimmte Architektur sind. Zum Beispiel ist [BERT](https://huggingface.co/bert-base-uncased) eine Architektur, wรคhrend `bert-base-uncased` ein Checkpoint ist. Modell ist ein allgemeiner Begriff, der entweder Architektur oder Prรผfpunkt bedeuten kann.
</Tip>
In dieser Anleitung lernen Sie, wie man:
* Einen vortrainierten Tokenizer lรคdt.
* Einen vortrainierten Merkmalsextraktor lรคdt.
* Einen vortrainierten Prozessor lรคdt.
* Ein vortrainiertes Modell lรคdt.
## AutoTokenizer
Nahezu jede NLP-Aufgabe beginnt mit einem Tokenizer. Ein Tokenizer wandelt Ihre Eingabe in ein Format um, das vom Modell verarbeitet werden kann.
Laden Sie einen Tokenizer mit [`AutoTokenizer.from_pretrained`]:
```py
>>> from transformers import AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("bert-base-uncased")
```
Dann tokenisieren Sie Ihre Eingabe wie unten gezeigt:
```py
>>> sequence = "In a hole in the ground there lived a hobbit."
>>> print(tokenizer(sequence))
{'input_ids': [101, 1999, 1037, 4920, 1999, 1996, 2598, 2045, 2973, 1037, 7570, 10322, 4183, 1012, 102],
'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}
```
## AutoFeatureExtractor
Fรผr Audio- und Bildverarbeitungsaufgaben verarbeitet ein Merkmalsextraktor das Audiosignal oder Bild in das richtige Eingabeformat.
Laden Sie einen Merkmalsextraktor mit [`AutoFeatureExtractor.from_pretrained`]:
```py
>>> from transformers import AutoFeatureExtractor
>>> feature_extractor = AutoFeatureExtractor.from_pretrained(
... "ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition"
... )
```
## AutoProcessor
Multimodale Aufgaben erfordern einen Prozessor, der zwei Arten von Vorverarbeitungswerkzeugen kombiniert. Das Modell [LayoutLMV2](model_doc/layoutlmv2) beispielsweise benรถtigt einen Feature-Extraktor fรผr Bilder und einen Tokenizer fรผr Text; ein Prozessor kombiniert beide.
Laden Sie einen Prozessor mit [`AutoProcessor.from_pretrained`]:
```py
>>> from transformers import AutoProcessor
>>> processor = AutoProcessor.from_pretrained("microsoft/layoutlmv2-base-uncased")
```
## AutoModel
<frameworkcontent>
<pt>
Mit den `AutoModelFor`-Klassen kรถnnen Sie schlieรlich ein vortrainiertes Modell fรผr eine bestimmte Aufgabe laden (siehe [hier](model_doc/auto) fรผr eine vollstรคndige Liste der verfรผgbaren Aufgaben). Laden Sie zum Beispiel ein Modell fรผr die Sequenzklassifikation mit [`AutoModelForSequenceClassification.from_pretrained`]:
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased")
```
Sie kรถnnen denselben Prรผfpunkt problemlos wiederverwenden, um eine Architektur fรผr eine andere Aufgabe zu laden:
```py
>>> from transformers import AutoModelForTokenClassification
>>> model = AutoModelForTokenClassification.from_pretrained("distilbert-base-uncased")
```
<Tip warning={true}>
Fรผr PyTorch-Modelle verwendet die Methode `from_pretrained()` `torch.load()`, die intern `pickle` verwendet und als unsicher bekannt ist. Generell sollte man niemals ein Modell laden, das aus einer nicht vertrauenswรผrdigen Quelle stammen kรถnnte, oder das manipuliert worden sein kรถnnte. Dieses Sicherheitsrisiko wird fรผr รถffentliche Modelle, die auf dem Hugging Face Hub gehostet werden, teilweise gemildert, da diese bei jeder รbertragung [auf Malware](https://huggingface.co/docs/hub/security-malware) gescannt werden. Siehe die [Hub-Dokumentation](https://huggingface.co/docs/hub/security) fรผr Best Practices wie [signierte Commit-Verifizierung](https://huggingface.co/docs/hub/security-gpg#signing-commits-with-gpg) mit GPG.
TensorFlow- und Flax-Checkpoints sind nicht betroffen und kรถnnen in PyTorch-Architekturen mit den Kwargs `from_tf` und `from_flax` fรผr die Methode `from_pretrained` geladen werden, um dieses Problem zu umgehen.
</Tip>
Im Allgemeinen empfehlen wir die Verwendung der Klasse "AutoTokenizer" und der Klasse "AutoModelFor", um trainierte Instanzen von Modellen zu laden. Dadurch wird sichergestellt, dass Sie jedes Mal die richtige Architektur laden. Im nรคchsten [Tutorial] (Vorverarbeitung) erfahren Sie, wie Sie Ihren neu geladenen Tokenizer, Feature Extractor und Prozessor verwenden, um einen Datensatz fรผr die Feinabstimmung vorzuverarbeiten.
</pt>
<tf>
Mit den Klassen `TFAutoModelFor` schlieรlich kรถnnen Sie ein vortrainiertes Modell fรผr eine bestimmte Aufgabe laden (siehe [hier](model_doc/auto) fรผr eine vollstรคndige Liste der verfรผgbaren Aufgaben). Laden Sie zum Beispiel ein Modell fรผr die Sequenzklassifikation mit [`TFAutoModelForSequenceClassification.from_pretrained`]:
```py
>>> from transformers import TFAutoModelForSequenceClassification
>>> model = TFAutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased")
```
Sie kรถnnen denselben Prรผfpunkt problemlos wiederverwenden, um eine Architektur fรผr eine andere Aufgabe zu laden:
```py
>>> from transformers import TFAutoModelForTokenClassification
>>> model = TFAutoModelForTokenClassification.from_pretrained("distilbert-base-uncased")
```
Im Allgemeinen empfehlen wir, die Klasse "AutoTokenizer" und die Klasse "TFAutoModelFor" zu verwenden, um vortrainierte Instanzen von Modellen zu laden. Dadurch wird sichergestellt, dass Sie jedes Mal die richtige Architektur laden. Im nรคchsten [Tutorial] (Vorverarbeitung) erfahren Sie, wie Sie Ihren neu geladenen Tokenizer, Feature Extractor und Prozessor verwenden, um einen Datensatz fรผr die Feinabstimmung vorzuverarbeiten.
</tf>
</frameworkcontent>
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/peft.md
|
<!--Copyright 2023 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Adapter mit ๐ค PEFT laden
[[open-in-colab]]
Die [Parameter-Efficient Fine Tuning (PEFT)](https://huggingface.co/blog/peft) Methoden frieren die vorab trainierten Modellparameter wรคhrend der Feinabstimmung ein und fรผgen eine kleine Anzahl trainierbarer Parameter (die Adapter) hinzu. Die Adapter werden trainiert, um aufgabenspezifische Informationen zu lernen. Es hat sich gezeigt, dass dieser Ansatz sehr speichereffizient ist und weniger Rechenleistung beansprucht, wรคhrend die Ergebnisse mit denen eines vollstรคndig feinabgestimmten Modells vergleichbar sind.
Adapter, die mit PEFT trainiert wurden, sind in der Regel um eine Grรถรenordnung kleiner als das vollstรคndige Modell, so dass sie bequem gemeinsam genutzt, gespeichert und geladen werden kรถnnen.
<div class="flex flex-col justify-center">
<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/peft/PEFT-hub-screenshot.png"/>
<figcaption class="text-center">Die Adaptergewichte fรผr ein OPTForCausalLM-Modell, die auf dem Hub gespeichert sind, sind nur ~6MB groร, verglichen mit der vollen Grรถรe der Modellgewichte, die ~700MB betragen kรถnnen.</figcaption>
</div>
Wenn Sie mehr รผber die ๐ค PEFT-Bibliothek erfahren mรถchten, sehen Sie sich die [Dokumentation](https://huggingface.co/docs/peft/index) an.
## Setup
Starten Sie mit der Installation von ๐ค PEFT:
```bash
pip install peft
```
Wenn Sie die brandneuen Funktionen ausprobieren mรถchten, sollten Sie die Bibliothek aus dem Quellcode installieren:
```bash
pip install git+https://github.com/huggingface/peft.git
```
## Unterstรผtzte PEFT-Modelle
Transformers unterstรผtzt nativ einige PEFT-Methoden, d.h. Sie kรถnnen lokal oder auf dem Hub gespeicherte Adaptergewichte laden und sie mit wenigen Zeilen Code einfach ausfรผhren oder trainieren. Die folgenden Methoden werden unterstรผtzt:
- [Low Rank Adapters](https://huggingface.co/docs/peft/conceptual_guides/lora)
- [IA3](https://huggingface.co/docs/peft/conceptual_guides/ia3)
- [AdaLoRA](https://arxiv.org/abs/2303.10512)
Wenn Sie andere PEFT-Methoden, wie z.B. Prompt Learning oder Prompt Tuning, verwenden mรถchten, oder รผber die ๐ค PEFT-Bibliothek im Allgemeinen, lesen Sie bitte die [Dokumentation](https://huggingface.co/docs/peft/index).
## Laden Sie einen PEFT-Adapter
Um ein PEFT-Adaptermodell von ๐ค Transformers zu laden und zu verwenden, stellen Sie sicher, dass das Hub-Repository oder das lokale Verzeichnis eine `adapter_config.json`-Datei und die Adaptergewichte enthรคlt, wie im obigen Beispielbild gezeigt. Dann kรถnnen Sie das PEFT-Adaptermodell mit der Klasse `AutoModelFor` laden. Um zum Beispiel ein PEFT-Adaptermodell fรผr die kausale Sprachmodellierung zu laden:
1. Geben Sie die PEFT-Modell-ID an.
2. รผbergeben Sie es an die Klasse [`AutoModelForCausalLM`].
```py
from transformers import AutoModelForCausalLM, AutoTokenizer
peft_model_id = "ybelkada/opt-350m-lora"
model = AutoModelForCausalLM.from_pretrained(peft_model_id)
```
<Tip>
Sie kรถnnen einen PEFT-Adapter entweder mit einer `AutoModelFor`-Klasse oder der Basismodellklasse wie `OPTForCausalLM` oder `LlamaForCausalLM` laden.
</Tip>
Sie kรถnnen einen PEFT-Adapter auch laden, indem Sie die Methode `load_adapter` aufrufen:
```py
from transformers import AutoModelForCausalLM, AutoTokenizer
model_id = "facebook/opt-350m"
peft_model_id = "ybelkada/opt-350m-lora"
model = AutoModelForCausalLM.from_pretrained(model_id)
model.load_adapter(peft_model_id)
```
## Laden in 8bit oder 4bit
Die `bitsandbytes`-Integration unterstรผtzt Datentypen mit 8bit und 4bit Genauigkeit, was fรผr das Laden groรer Modelle nรผtzlich ist, weil es Speicher spart (lesen Sie den `bitsandbytes`-Integrations [guide](./quantization#bitsandbytes-integration), um mehr zu erfahren). Fรผgen Sie die Parameter `load_in_8bit` oder `load_in_4bit` zu [`~PreTrainedModel.from_pretrained`] hinzu und setzen Sie `device_map="auto"`, um das Modell effektiv auf Ihre Hardware zu verteilen:
```py
from transformers import AutoModelForCausalLM, AutoTokenizer
peft_model_id = "ybelkada/opt-350m-lora"
model = AutoModelForCausalLM.from_pretrained(peft_model_id, device_map="auto", load_in_8bit=True)
```
## Einen neuen Adapter hinzufรผgen
Sie kรถnnen [`~peft.PeftModel.add_adapter`] verwenden, um einen neuen Adapter zu einem Modell mit einem bestehenden Adapter hinzuzufรผgen, solange der neue Adapter vom gleichen Typ ist wie der aktuelle Adapter. Wenn Sie zum Beispiel einen bestehenden LoRA-Adapter an ein Modell angehรคngt haben:
```py
from transformers import AutoModelForCausalLM, OPTForCausalLM, AutoTokenizer
from peft import PeftConfig
model_id = "facebook/opt-350m"
model = AutoModelForCausalLM.from_pretrained(model_id)
lora_config = LoraConfig(
target_modules=["q_proj", "k_proj"],
init_lora_weights=False
)
model.add_adapter(lora_config, adapter_name="adapter_1")
```
Um einen neuen Adapter hinzuzufรผgen:
```py
# attach new adapter with same config
model.add_adapter(lora_config, adapter_name="adapter_2")
```
Jetzt kรถnnen Sie mit [`~peft.PeftModel.set_adapter`] festlegen, welcher Adapter verwendet werden soll:
```py
# use adapter_1
model.set_adapter("adapter_1")
output = model.generate(**inputs)
print(tokenizer.decode(output_disabled[0], skip_special_tokens=True))
# use adapter_2
model.set_adapter("adapter_2")
output_enabled = model.generate(**inputs)
print(tokenizer.decode(output_enabled[0], skip_special_tokens=True))
```
## Aktivieren und Deaktivieren von Adaptern
Sobald Sie einen Adapter zu einem Modell hinzugefรผgt haben, kรถnnen Sie das Adaptermodul aktivieren oder deaktivieren. So aktivieren Sie das Adaptermodul:
```py
from transformers import AutoModelForCausalLM, OPTForCausalLM, AutoTokenizer
from peft import PeftConfig
model_id = "facebook/opt-350m"
adapter_model_id = "ybelkada/opt-350m-lora"
tokenizer = AutoTokenizer.from_pretrained(model_id)
text = "Hello"
inputs = tokenizer(text, return_tensors="pt")
model = AutoModelForCausalLM.from_pretrained(model_id)
peft_config = PeftConfig.from_pretrained(adapter_model_id)
# to initiate with random weights
peft_config.init_lora_weights = False
model.add_adapter(peft_config)
model.enable_adapters()
output = model.generate(**inputs)
```
So deaktivieren Sie das Adaptermodul:
```py
model.disable_adapters()
output = model.generate(**inputs)
```
## PEFT-Adapter trainieren
PEFT-Adapter werden von der Klasse [`Trainer`] unterstรผtzt, so dass Sie einen Adapter fรผr Ihren speziellen Anwendungsfall trainieren kรถnnen. Dazu mรผssen Sie nur ein paar weitere Codezeilen hinzufรผgen. Zum Beispiel, um einen LoRA-Adapter zu trainieren:
<Tip>
Wenn Sie mit der Feinabstimmung eines Modells mit [`Trainer`] noch nicht vertraut sind, werfen Sie einen Blick auf das Tutorial [Feinabstimmung eines vortrainierten Modells](Training).
</Tip>
1. Definieren Sie Ihre Adapterkonfiguration mit dem Aufgabentyp und den Hyperparametern (siehe [`~peft.LoraConfig`] fรผr weitere Details darรผber, was die Hyperparameter tun).
```py
from peft import LoraConfig
peft_config = LoraConfig(
lora_alpha=16,
lora_dropout=0.1,
r=64,
bias="none",
task_type="CAUSAL_LM",
)
```
2. Fรผgen Sie dem Modell einen Adapter hinzu.
```py
model.add_adapter(peft_config)
```
3. Jetzt kรถnnen Sie das Modell an [`Trainer`] รผbergeben!
```py
trainer = Trainer(model=model, ...)
trainer.train()
```
So speichern Sie Ihren trainierten Adapter und laden ihn wieder:
```py
model.save_pretrained(save_dir)
model = AutoModelForCausalLM.from_pretrained(save_dir)
```
<!--
TODO: (@younesbelkada @stevhliu)
- Link to PEFT docs for further details
- Trainer
- 8-bit / 4-bit examples ?
-->
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/accelerate.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Verteiltes Training mit ๐ค Accelerate
Da die Modelle immer grรถรer werden, hat sich die Parallelitรคt als Strategie zum Trainieren grรถรerer Modelle auf begrenzter Hardware und zur Beschleunigung der Trainingsgeschwindigkeit um mehrere Grรถรenordnungen erwiesen. Bei Hugging Face haben wir die Bibliothek [๐ค Accelerate](https://huggingface.co/docs/accelerate) entwickelt, um Nutzern zu helfen, ein ๐ค Transformers-Modell auf jeder Art von verteiltem Setup zu trainieren, egal ob es sich um mehrere GPUs auf einer Maschine oder mehrere GPUs auf mehreren Maschinen handelt. In diesem Tutorial lernen Sie, wie Sie Ihre native PyTorch-Trainingsschleife anpassen, um das Training in einer verteilten Umgebung zu ermรถglichen.
## Einrichtung
Beginnen Sie mit der Installation von ๐ค Accelerate:
```bash
pip install accelerate
```
Dann importieren und erstellen Sie ein [`~accelerate.Accelerator`]-Objekt. Der [`~accelerate.Accelerator`] wird automatisch Ihre Art der verteilten Einrichtung erkennen und alle notwendigen Komponenten fรผr das Training initialisieren. Sie mรผssen Ihr Modell nicht explizit auf einem Gerรคt platzieren.
```py
>>> from accelerate import Accelerator
>>> accelerator = Accelerator()
```
## Vorbereiten auf die Beschleunigung
Der nรคchste Schritt ist die รbergabe aller relevanten Trainingsobjekte an die Methode [`~accelerate.Accelerator.prepare`]. Dazu gehรถren Ihre Trainings- und Evaluierungs-DataLoader, ein Modell und ein Optimierer:
```py
>>> train_dataloader, eval_dataloader, model, optimizer = accelerator.prepare(
... train_dataloader, eval_dataloader, model, optimizer
... )
```
## Rรผckwรคrts
Die letzte Ergรคnzung besteht darin, das typische `loss.backward()` in der Trainingsschleife durch die ๐ค Accelerate-Methode [`~accelerate.Accelerator.backward`] zu ersetzen:
```py
>>> for epoch in range(num_epochs):
... for batch in train_dataloader:
... outputs = model(**batch)
... loss = outputs.loss
... accelerator.backward(loss)
... optimizer.step()
... lr_scheduler.step()
... optimizer.zero_grad()
... progress_bar.update(1)
```
Wie Sie im folgenden Code sehen kรถnnen, mรผssen Sie nur vier zusรคtzliche Codezeilen zu Ihrer Trainingsschleife hinzufรผgen, um verteiltes Training zu ermรถglichen!
```diff
+ from accelerate import Accelerator
from transformers import AdamW, AutoModelForSequenceClassification, get_scheduler
+ accelerator = Accelerator()
model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2)
optimizer = AdamW(model.parameters(), lr=3e-5)
- device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
- model.to(device)
+ train_dataloader, eval_dataloader, model, optimizer = accelerator.prepare(
+ train_dataloader, eval_dataloader, model, optimizer
+ )
num_epochs = 3
num_training_steps = num_epochs * len(train_dataloader)
lr_scheduler = get_scheduler(
"linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=num_training_steps
)
progress_bar = tqdm(range(num_training_steps))
model.train()
for epoch in range(num_epochs):
for batch in train_dataloader:
- batch = {k: v.to(device) for k, v in batch.items()}
outputs = model(**batch)
loss = outputs.loss
- loss.backward()
+ accelerator.backward(loss)
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
progress_bar.update(1)
```
## Trainieren
Sobald Sie die entsprechenden Codezeilen hinzugefรผgt haben, starten Sie Ihr Training in einem Skript oder einem Notebook wie Colaboratory.
### Trainieren mit einem Skript
Wenn Sie Ihr Training mit einem Skript durchfรผhren, fรผhren Sie den folgenden Befehl aus, um eine Konfigurationsdatei zu erstellen und zu speichern:
```bash
accelerate config
```
Dann starten Sie Ihr Training mit:
```bash
accelerate launch train.py
```
### Trainieren mit einem Notebook
๐ค Accelerate kann auch in einem Notebook laufen, wenn Sie planen, die TPUs von Colaboratory zu verwenden. Verpacken Sie den gesamten Code, der fรผr das Training verantwortlich ist, in eine Funktion und รผbergeben Sie diese an [`~accelerate.notebook_launcher`]:
```py
>>> from accelerate import notebook_launcher
>>> notebook_launcher(training_function)
```
Weitere Informationen รผber ๐ค Accelerate und seine umfangreichen Funktionen finden Sie in der [Dokumentation](https://huggingface.co/docs/accelerate).
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/pr_checks.md
|
<!---
Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# รberprรผfungen bei einer Pull-Anfrage
Wenn Sie eine Pull-Anfrage fรผr ๐ค Transformers รถffnen, wird eine ganze Reihe von Prรผfungen durchgefรผhrt, um sicherzustellen, dass der Patch, den Sie hinzufรผgen, nichts Bestehendes zerstรถrt. Es gibt vier Arten von Prรผfungen:
- regulรคre Tests
- Erstellung der Dokumentation
- Stil von Code und Dokumentation
- allgemeine Konsistenz des Repository
In diesem Dokument werden wir versuchen zu erklรคren, worum es sich bei diesen verschiedenen Prรผfungen handelt und wie Sie sie lokal debuggen kรถnnen, wenn eine der Prรผfungen in Ihrer PR fehlschlรคgt.
Beachten Sie, dass Sie im Idealfall eine Dev-Installation benรถtigen:
```bash
pip install transformers[dev]
```
oder fรผr eine bearbeitbare Installation:
```bash
pip install -e .[dev]
```
innerhalb des Transformers Repo. Da die Anzahl der optionalen Abhรคngigkeiten von Transformers stark zugenommen hat, ist es mรถglich, dass Sie nicht alle davon bekommen kรถnnen. Wenn die Dev-Installation fehlschlรคgt, stellen Sie sicher, dass Sie das Deep Learning-Framework, mit dem Sie arbeiten, installieren (PyTorch, TensorFlow und/oder Flax).
```bash
pip install transformers[quality]
```
oder fรผr eine bearbeitbare Installation:
```bash
pip install -e .[quality]
```
## Tests
Alle Jobs, die mit `ci/circleci: run_tests_` beginnen, fรผhren Teile der Transformers-Testsuite aus. Jeder dieser Jobs konzentriert sich auf einen Teil der Bibliothek in einer bestimmten Umgebung: `ci/circleci: run_tests_pipelines_tf` zum Beispiel fรผhrt den Pipelines-Test in einer Umgebung aus, in der nur TensorFlow installiert ist.
Beachten Sie, dass nur ein Teil der Testsuite jedes Mal ausgefรผhrt wird, um zu vermeiden, dass Tests ausgefรผhrt werden, wenn es keine wirkliche รnderung in den Modulen gibt, die sie testen: ein Dienstprogramm wird ausgefรผhrt, um die Unterschiede in der Bibliothek zwischen vor und nach dem PR zu ermitteln (was GitHub Ihnen auf der Registerkarte "Files changes" anzeigt) und die Tests auszuwรคhlen, die von diesem Unterschied betroffen sind. Dieses Dienstprogramm kann lokal mit ausgefรผhrt werden:
```bash
python utils/tests_fetcher.py
```
aus dem Stammverzeichnis des Transformers-Repositoriums. Es wird:
1. รberprรผfen Sie fรผr jede Datei im Diff, ob die รnderungen im Code oder nur in Kommentaren oder Docstrings enthalten sind. Nur die Dateien mit echten Codeรคnderungen werden beibehalten.
2. Erstellen Sie eine interne Map, die fรผr jede Datei des Quellcodes der Bibliothek alle Dateien angibt, auf die sie rekursiv Einfluss nimmt. Von Modul A wird gesagt, dass es sich auf Modul B auswirkt, wenn Modul B Modul A importiert. Fรผr die rekursive Auswirkung benรถtigen wir eine Kette von Modulen, die von Modul A zu Modul B fรผhrt und in der jedes Modul das vorherige importiert.
3. Wenden Sie diese Zuordnung auf die in Schritt 1 gesammelten Dateien an. So erhalten wir die Liste der Modelldateien, die von der PR betroffen sind.
4. Ordnen Sie jede dieser Dateien der/den entsprechenden Testdatei(en) zu und erhalten Sie die Liste der auszufรผhrenden Tests.
Wenn Sie das Skript lokal ausfรผhren, sollten Sie die Ergebnisse von Schritt 1, 3 und 4 ausgegeben bekommen und somit wissen, welche Tests ausgefรผhrt werden. Das Skript erstellt auรerdem eine Datei namens `test_list.txt`, die die Liste der auszufรผhrenden Tests enthรคlt, die Sie mit dem folgenden Befehl lokal ausfรผhren kรถnnen:
```bash
python -m pytest -n 8 --dist=loadfile -rA -s $(cat test_list.txt)
```
Fรผr den Fall, dass Ihnen etwas entgangen ist, wird die komplette Testreihe ebenfalls tรคglich ausgefรผhrt.
## Dokumentation erstellen
Der Job `build_pr_documentation` erstellt und generiert eine Vorschau der Dokumentation, um sicherzustellen, dass alles in Ordnung ist, wenn Ihr PR zusammengefรผhrt wird. Ein Bot fรผgt einen Link zur Vorschau der Dokumentation zu Ihrem PR hinzu. Alle รnderungen, die Sie an dem PR vornehmen, werden automatisch in der Vorschau aktualisiert. Wenn die Dokumentation nicht erstellt werden kann, klicken Sie auf **Details** neben dem fehlgeschlagenen Auftrag, um zu sehen, wo der Fehler liegt. Oft ist der Fehler so einfach wie eine fehlende Datei im `toctree`.
Wenn Sie daran interessiert sind, die Dokumentation lokal zu erstellen oder in der Vorschau anzusehen, werfen Sie einen Blick in die [`README.md`](https://github.com/huggingface/transformers/tree/main/docs) im Ordner docs.
## Code und Dokumentationsstil
Die Formatierung des Codes erfolgt fรผr alle Quelldateien, die Beispiele und die Tests mit `black` und `ruff`. Wir haben auch ein benutzerdefiniertes Tool, das sich um die Formatierung von docstrings und `rst`-Dateien kรผmmert (`utils/style_doc.py`), sowie um die Reihenfolge der Lazy-Importe, die in den Transformers `__init__.py`-Dateien durchgefรผhrt werden (`utils/custom_init_isort.py`). All dies kรถnnen Sie starten, indem Sie Folgendes ausfรผhren
```bash
make style
```
Das CI prรผft, ob diese innerhalb der Prรผfung `ci/circleci: check_code_quality` angewendet wurden. Es fรผhrt auch `ruff` aus, das einen grundlegenden Blick auf Ihren Code wirft und sich beschwert, wenn es eine undefinierte Variable findet oder eine, die nicht verwendet wird. Um diese Prรผfung lokal auszufรผhren, verwenden Sie
```bash
make quality
```
Dies kann sehr viel Zeit in Anspruch nehmen. Um dasselbe nur fรผr die Dateien zu tun, die Sie im aktuellen Zweig geรคndert haben, fรผhren Sie
```bash
make fixup
```
Dieser letzte Befehl fรผhrt auch alle zusรคtzlichen Prรผfungen fรผr die Konsistenz des Repositorys durch. Schauen wir uns diese an.
## Repository-Konsistenz
Dies fasst alle Tests zusammen, die sicherstellen, dass Ihr PR das Repository in einem guten Zustand verlรคsst. Sie kรถnnen diese Prรผfung lokal durchfรผhren, indem Sie Folgendes ausfรผhren:
```bash
make repo-consistency
```
Dies รผberprรผft, ob:
- Alle zum Init hinzugefรผgten Objekte sind dokumentiert (ausgefรผhrt von `utils/check_repo.py`)
- Alle `__init__.py`-Dateien haben in ihren beiden Abschnitten den gleichen Inhalt (ausgefรผhrt von `utils/check_inits.py`)
- Der gesamte Code, der als Kopie eines anderen Moduls identifiziert wurde, stimmt mit dem Original รผberein (ausgefรผhrt von `utils/check_copies.py`)
- Alle Konfigurationsklassen haben mindestens einen gรผltigen Prรผfpunkt, der in ihren Dokumentationen erwรคhnt wird (ausgefรผhrt von `utils/check_config_docstrings.py`)
- Alle Konfigurationsklassen enthalten nur Attribute, die in den entsprechenden Modellierungsdateien verwendet werden (ausgefรผhrt von `utils/check_config_attributes.py`)
- Die รbersetzungen der READMEs und der Index des Dokuments haben die gleiche Modellliste wie die Haupt-README (durchgefรผhrt von `utils/check_copies.py`)
- Die automatisch generierten Tabellen in der Dokumentation sind auf dem neuesten Stand (ausgefรผhrt von `utils/check_table.py`)
- Die Bibliothek verfรผgt รผber alle Objekte, auch wenn nicht alle optionalen Abhรคngigkeiten installiert sind (ausgefรผhrt von `utils/check_dummies.py`)
Sollte diese Prรผfung fehlschlagen, mรผssen die ersten beiden Punkte manuell korrigiert werden, die letzten vier kรถnnen automatisch fรผr Sie korrigiert werden, indem Sie den Befehl
```bash
make fix-copies
```
Zusรคtzliche Prรผfungen betreffen PRs, die neue Modelle hinzufรผgen, vor allem, dass:
- Alle hinzugefรผgten Modelle befinden sich in einer Auto-Zuordnung (durchgefรผhrt von `utils/check_repo.py`)
<!-- TODO Sylvain, add a check that makes sure the common tests are implemented.-->
- Alle Modelle werden ordnungsgemรคร getestet (ausgefรผhrt von `utils/check_repo.py`)
<!-- TODO Sylvain, add the following
- All models are added to the main README, inside the main doc
- All checkpoints used actually exist on the Hub
-->
### Kopien prรผfen
Da die Transformers-Bibliothek in Bezug auf den Modellcode sehr eigenwillig ist und jedes Modell vollstรคndig in einer einzigen Datei implementiert sein sollte, ohne sich auf andere Modelle zu stรผtzen, haben wir einen Mechanismus hinzugefรผgt, der รผberprรผft, ob eine Kopie des Codes einer Ebene eines bestimmten Modells mit dem Original รผbereinstimmt. Auf diese Weise kรถnnen wir bei einer Fehlerbehebung alle anderen betroffenen Modelle sehen und entscheiden, ob wir die รnderung weitergeben oder die Kopie zerstรถren.
<Tip>
Wenn eine Datei eine vollstรคndige Kopie einer anderen Datei ist, sollten Sie sie in der Konstante `FULL_COPIES` von `utils/check_copies.py` registrieren.
</Tip>
Dieser Mechanismus stรผtzt sich auf Kommentare der Form `# Kopiert von xxx`. Das `xxx` sollte den gesamten Pfad zu der Klasse der Funktion enthalten, die darunter kopiert wird. Zum Beispiel ist `RobertaSelfOutput` eine direkte Kopie der Klasse `BertSelfOutput`. Sie kรถnnen also [hier](https://github.com/huggingface/transformers/blob/2bd7a27a671fd1d98059124024f580f8f5c0f3b5/src/transformers/models/roberta/modeling_roberta.py#L289) sehen, dass sie einen Kommentar hat:
```py
# Copied from transformers.models.bert.modeling_bert.BertSelfOutput
```
Beachten Sie, dass Sie dies nicht auf eine ganze Klasse anwenden, sondern auf die entsprechenden Methoden, von denen kopiert wird. Zum Beispiel [hier](https://github.com/huggingface/transformers/blob/2bd7a27a671fd1d98059124024f580f8f5c0f3b5/src/transformers/models/roberta/modeling_roberta.py#L598) kรถnnen Sie sehen, wie `RobertaPreTrainedModel._init_weights` von der gleichen Methode in `BertPreTrainedModel` mit dem Kommentar kopiert wird:
```py
# Copied from transformers.models.bert.modeling_bert.BertPreTrainedModel._init_weights
```
Manchmal ist die Kopie bis auf die Namen genau gleich: zum Beispiel verwenden wir in `RobertaAttention` `RobertaSelfAttention` anstelle von `BertSelfAttention`, aber ansonsten ist der Code genau derselbe. Aus diesem Grund unterstรผtzt `#Copied from` einfache String-Ersetzungen mit der folgenden Syntax: `Kopiert von xxx mit foo->bar`. Das bedeutet, dass der Code kopiert wird, wobei alle Instanzen von "foo" durch "bar" ersetzt werden. Sie kรถnnen sehen, wie es [hier](https://github.com/huggingface/transformers/blob/2bd7a27a671fd1d98059124024f580f8f5c0f3b5/src/transformers/models/roberta/modeling_roberta.py#L304C1-L304C86) in `RobertaAttention` mit dem Kommentar verwendet wird:
```py
# Copied from transformers.models.bert.modeling_bert.BertAttention with Bert->Roberta
```
Beachten Sie, dass um den Pfeil herum keine Leerzeichen stehen sollten (es sei denn, das Leerzeichen ist Teil des zu ersetzenden Musters, natรผrlich).
Sie kรถnnen mehrere Muster durch ein Komma getrennt hinzufรผgen. Zum Beispiel ist hier `CamemberForMaskedLM` eine direkte Kopie von `RobertaForMaskedLM` mit zwei Ersetzungen: `Roberta` zu `Camembert` und `ROBERTA` zu `CAMEMBERT`. Sie kรถnnen [hier](https://github.com/huggingface/transformers/blob/15082a9dc6950ecae63a0d3e5060b2fc7f15050a/src/transformers/models/camembert/modeling_camembert.py#L929) sehen, wie dies mit dem Kommentar gemacht wird:
```py
# Copied from transformers.models.roberta.modeling_roberta.RobertaForMaskedLM with Roberta->Camembert, ROBERTA->CAMEMBERT
```
Wenn die Reihenfolge eine Rolle spielt (weil eine der Ersetzungen mit einer vorherigen in Konflikt geraten kรถnnte), werden die Ersetzungen von links nach rechts ausgefรผhrt.
<Tip>
Wenn die Ersetzungen die Formatierung รคndern (wenn Sie z.B. einen kurzen Namen durch einen sehr langen Namen ersetzen), wird die Kopie nach Anwendung des automatischen Formats รผberprรผft.
</Tip>
Eine andere Mรถglichkeit, wenn es sich bei den Mustern nur um verschiedene Umschreibungen derselben Ersetzung handelt (mit einer groร- und einer kleingeschriebenen Variante), besteht darin, die Option `all-casing` hinzuzufรผgen. [Hier](https://github.com/huggingface/transformers/blob/15082a9dc6950ecae63a0d3e5060b2fc7f15050a/src/transformers/models/mobilebert/modeling_mobilebert.py#L1237) ist ein Beispiel in `MobileBertForSequenceClassification` mit dem Kommentar:
```py
# Copied from transformers.models.bert.modeling_bert.BertForSequenceClassification with Bert->MobileBert all-casing
```
In diesem Fall wird der Code von `BertForSequenceClassification` kopiert, indem er ersetzt wird:
- `Bert` durch `MobileBert` (zum Beispiel bei der Verwendung von `MobileBertModel` in der Init)
- `bert` durch `mobilebert` (zum Beispiel bei der Definition von `self.mobilebert`)
- `BERT` durch `MOBILEBERT` (in der Konstante `MOBILEBERT_INPUTS_DOCSTRING`)
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/add_new_pipeline.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Wie erstellt man eine benutzerdefinierte Pipeline?
In dieser Anleitung sehen wir uns an, wie Sie eine benutzerdefinierte Pipeline erstellen und sie auf dem [Hub](hf.co/models) freigeben oder sie der
๐ค Transformers-Bibliothek hinzufรผgen.
Zuallererst mรผssen Sie entscheiden, welche Roheingaben die Pipeline verarbeiten kann. Es kann sich um Strings, rohe Bytes,
Dictionaries oder was auch immer die wahrscheinlichste gewรผnschte Eingabe ist. Versuchen Sie, diese Eingaben so rein wie mรถglich in Python zu halten
denn das macht die Kompatibilitรคt einfacher (auch mit anderen Sprachen รผber JSON). Dies werden die Eingaben der
Pipeline (`Vorverarbeitung`).
Definieren Sie dann die `Outputs`. Dieselbe Richtlinie wie fรผr die Eingรคnge. Je einfacher, desto besser. Dies werden die Ausgaben der
Methode `Postprocess`.
Beginnen Sie damit, die Basisklasse `Pipeline` mit den 4 Methoden zu erben, die fรผr die Implementierung von `preprocess` benรถtigt werden,
Weiterleiten", "Nachbearbeitung" und "Parameter sรคubern".
```python
from transformers import Pipeline
class MyPipeline(Pipeline):
def _sanitize_parameters(self, **kwargs):
preprocess_kwargs = {}
if "maybe_arg" in kwargs:
preprocess_kwargs["maybe_arg"] = kwargs["maybe_arg"]
return preprocess_kwargs, {}, {}
def preprocess(self, inputs, maybe_arg=2):
model_input = Tensor(inputs["input_ids"])
return {"model_input": model_input}
def _forward(self, model_inputs):
# model_inputs == {"model_input": model_input}
outputs = self.model(**model_inputs)
# Maybe {"logits": Tensor(...)}
return outputs
def postprocess(self, model_outputs):
best_class = model_outputs["logits"].softmax(-1)
return best_class
```
Die Struktur dieser Aufteilung soll eine relativ nahtlose Unterstรผtzung fรผr CPU/GPU ermรถglichen und gleichzeitig die Durchfรผhrung von
Vor-/Nachbearbeitung auf der CPU in verschiedenen Threads
Preprocess" nimmt die ursprรผnglich definierten Eingaben und wandelt sie in etwas um, das in das Modell eingespeist werden kann. Es kann
mehr Informationen enthalten und ist normalerweise ein `Dict`.
`_forward` ist das Implementierungsdetail und ist nicht dafรผr gedacht, direkt aufgerufen zu werden. Weiterleiten" ist die bevorzugte
aufgerufene Methode, da sie Sicherheitsvorkehrungen enthรคlt, die sicherstellen, dass alles auf dem erwarteten Gerรคt funktioniert. Wenn etwas
mit einem realen Modell verknรผpft ist, gehรถrt es in die Methode `_forward`, alles andere gehรถrt in die Methoden preprocess/postprocess.
Die Methode `Postprocess` nimmt die Ausgabe von `_forward` und verwandelt sie in die endgรผltige Ausgabe, die zuvor festgelegt wurde.
zuvor entschieden wurde.
Die Methode `_sanitize_parameters` ermรถglicht es dem Benutzer, beliebige Parameter zu รผbergeben, wann immer er mรถchte, sei es bei der Initialisierung
Zeit `pipeline(...., maybe_arg=4)` oder zur Aufrufzeit `pipe = pipeline(...); output = pipe(...., maybe_arg=4)`.
Die Rรผckgabe von `_sanitize_parameters` sind die 3 Dicts von kwargs, die direkt an `preprocess` รผbergeben werden,
`_forward` und `postprocess` รผbergeben werden. Fรผllen Sie nichts aus, wenn der Aufrufer keinen zusรคtzlichen Parameter angegeben hat. Das
erlaubt es, die Standardargumente in der Funktionsdefinition beizubehalten, was immer "natรผrlicher" ist.
Ein klassisches Beispiel wรคre das Argument `top_k` in der Nachbearbeitung bei Klassifizierungsaufgaben.
```python
>>> pipe = pipeline("my-new-task")
>>> pipe("This is a test")
[{"label": "1-star", "score": 0.8}, {"label": "2-star", "score": 0.1}, {"label": "3-star", "score": 0.05}
{"label": "4-star", "score": 0.025}, {"label": "5-star", "score": 0.025}]
>>> pipe("This is a test", top_k=2)
[{"label": "1-star", "score": 0.8}, {"label": "2-star", "score": 0.1}]
```
In order to achieve that, we'll update our `postprocess` method with a default parameter to `5`. and edit
`_sanitize_parameters` to allow this new parameter.
```python
def postprocess(self, model_outputs, top_k=5):
best_class = model_outputs["logits"].softmax(-1)
# Add logic to handle top_k
return best_class
def _sanitize_parameters(self, **kwargs):
preprocess_kwargs = {}
if "maybe_arg" in kwargs:
preprocess_kwargs["maybe_arg"] = kwargs["maybe_arg"]
postprocess_kwargs = {}
if "top_k" in kwargs:
postprocess_kwargs["top_k"] = kwargs["top_k"]
return preprocess_kwargs, {}, postprocess_kwargs
```
Versuchen Sie, die Eingaben/Ausgaben sehr einfach und idealerweise JSON-serialisierbar zu halten, da dies die Verwendung der Pipeline sehr einfach macht
ohne dass die Benutzer neue Arten von Objekten verstehen mรผssen. Es ist auch relativ รผblich, viele verschiedene Arten von Argumenten zu unterstรผtzen
von Argumenten zu unterstรผtzen (Audiodateien, die Dateinamen, URLs oder reine Bytes sein kรถnnen).
## Hinzufรผgen zur Liste der unterstรผtzten Aufgaben
Um Ihre `neue Aufgabe` in die Liste der unterstรผtzten Aufgaben aufzunehmen, mรผssen Sie sie zur `PIPELINE_REGISTRY` hinzufรผgen:
```python
from transformers.pipelines import PIPELINE_REGISTRY
PIPELINE_REGISTRY.register_pipeline(
"new-task",
pipeline_class=MyPipeline,
pt_model=AutoModelForSequenceClassification,
)
```
Wenn Sie mรถchten, kรถnnen Sie ein Standardmodell angeben. In diesem Fall sollte es mit einer bestimmten Revision (die der Name einer Verzweigung oder ein Commit-Hash sein kann, hier haben wir `"abcdef"` genommen) sowie mit dem Typ versehen sein:
```python
PIPELINE_REGISTRY.register_pipeline(
"new-task",
pipeline_class=MyPipeline,
pt_model=AutoModelForSequenceClassification,
default={"pt": ("user/awesome_model", "abcdef")},
type="text", # current support type: text, audio, image, multimodal
)
```
## Teilen Sie Ihre Pipeline auf dem Hub
Um Ihre benutzerdefinierte Pipeline auf dem Hub freizugeben, mรผssen Sie lediglich den benutzerdefinierten Code Ihrer `Pipeline`-Unterklasse in einer
Python-Datei speichern. Nehmen wir zum Beispiel an, Sie mรถchten eine benutzerdefinierte Pipeline fรผr die Klassifizierung von Satzpaaren wie folgt verwenden:
```py
import numpy as np
from transformers import Pipeline
def softmax(outputs):
maxes = np.max(outputs, axis=-1, keepdims=True)
shifted_exp = np.exp(outputs - maxes)
return shifted_exp / shifted_exp.sum(axis=-1, keepdims=True)
class PairClassificationPipeline(Pipeline):
def _sanitize_parameters(self, **kwargs):
preprocess_kwargs = {}
if "second_text" in kwargs:
preprocess_kwargs["second_text"] = kwargs["second_text"]
return preprocess_kwargs, {}, {}
def preprocess(self, text, second_text=None):
return self.tokenizer(text, text_pair=second_text, return_tensors=self.framework)
def _forward(self, model_inputs):
return self.model(**model_inputs)
def postprocess(self, model_outputs):
logits = model_outputs.logits[0].numpy()
probabilities = softmax(logits)
best_class = np.argmax(probabilities)
label = self.model.config.id2label[best_class]
score = probabilities[best_class].item()
logits = logits.tolist()
return {"label": label, "score": score, "logits": logits}
```
Die Implementierung ist Framework-unabhรคngig und funktioniert fรผr PyTorch- und TensorFlow-Modelle. Wenn wir dies in einer Datei
einer Datei namens `pair_classification.py` gespeichert haben, kรถnnen wir sie importieren und wie folgt registrieren:
```py
from pair_classification import PairClassificationPipeline
from transformers.pipelines import PIPELINE_REGISTRY
from transformers import AutoModelForSequenceClassification, TFAutoModelForSequenceClassification
PIPELINE_REGISTRY.register_pipeline(
"pair-classification",
pipeline_class=PairClassificationPipeline,
pt_model=AutoModelForSequenceClassification,
tf_model=TFAutoModelForSequenceClassification,
)
```
Sobald dies geschehen ist, kรถnnen wir es mit einem vortrainierten Modell verwenden. Zum Beispiel wurde `sgugger/finetuned-bert-mrpc` auf den
auf den MRPC-Datensatz abgestimmt, der Satzpaare als Paraphrasen oder nicht klassifiziert.
```py
from transformers import pipeline
classifier = pipeline("pair-classification", model="sgugger/finetuned-bert-mrpc")
```
Dann kรถnnen wir sie auf dem Hub mit der Methode `save_pretrained` in einem `Repository` freigeben:
```py
from huggingface_hub import Repository
repo = Repository("test-dynamic-pipeline", clone_from="{your_username}/test-dynamic-pipeline")
classifier.save_pretrained("test-dynamic-pipeline")
repo.push_to_hub()
```
Dadurch wird die Datei, in der Sie `PairClassificationPipeline` definiert haben, in den Ordner `"test-dynamic-pipeline"` kopiert,
und speichert das Modell und den Tokenizer der Pipeline, bevor Sie alles in das Repository verschieben
`{Ihr_Benutzername}/test-dynamic-pipeline`. Danach kann jeder die Pipeline verwenden, solange er die Option
`trust_remote_code=True` angeben:
```py
from transformers import pipeline
classifier = pipeline(model="{your_username}/test-dynamic-pipeline", trust_remote_code=True)
```
## Hinzufรผgen der Pipeline zu ๐ค Transformers
Wenn Sie Ihre Pipeline zu ๐ค Transformers beitragen mรถchten, mรผssen Sie ein neues Modul im Untermodul `pipelines` hinzufรผgen
mit dem Code Ihrer Pipeline hinzufรผgen. Fรผgen Sie es dann der Liste der in `pipelines/__init__.py` definierten Aufgaben hinzu.
Dann mรผssen Sie noch Tests hinzufรผgen. Erstellen Sie eine neue Datei `tests/test_pipelines_MY_PIPELINE.py` mit Beispielen fรผr die anderen Tests.
Die Funktion `run_pipeline_test` ist sehr allgemein gehalten und lรคuft auf kleinen Zufallsmodellen auf jeder mรถglichen
Architektur, wie durch `model_mapping` und `tf_model_mapping` definiert.
Dies ist sehr wichtig, um die zukรผnftige Kompatibilitรคt zu testen, d.h. wenn jemand ein neues Modell fรผr
`XXXForQuestionAnswering` hinzufรผgt, wird der Pipeline-Test versuchen, mit diesem Modell zu arbeiten. Da die Modelle zufรคllig sind, ist es
ist es unmรถglich, die tatsรคchlichen Werte zu รผberprรผfen. Deshalb gibt es eine Hilfsfunktion `ANY`, die einfach versucht, die
Ausgabe der Pipeline TYPE.
Auรerdem *mรผssen* Sie 2 (idealerweise 4) Tests implementieren.
- test_small_model_pt` : Definieren Sie 1 kleines Modell fรผr diese Pipeline (es spielt keine Rolle, ob die Ergebnisse keinen Sinn ergeben)
und testen Sie die Ausgaben der Pipeline. Die Ergebnisse sollten die gleichen sein wie bei `test_small_model_tf`.
- test_small_model_tf : Definieren Sie 1 kleines Modell fรผr diese Pipeline (es spielt keine Rolle, ob die Ergebnisse keinen Sinn ergeben)
und testen Sie die Ausgaben der Pipeline. Die Ergebnisse sollten die gleichen sein wie bei `test_small_model_pt`.
- test_large_model_pt` (`optional`): Testet die Pipeline an einer echten Pipeline, bei der die Ergebnisse
Sinn machen. Diese Tests sind langsam und sollten als solche gekennzeichnet werden. Hier geht es darum, die Pipeline zu prรคsentieren und sicherzustellen
sicherzustellen, dass es in zukรผnftigen Versionen keine Abweichungen gibt.
- test_large_model_tf` (`optional`): Testet die Pipeline an einer echten Pipeline, bei der die Ergebnisse
Sinn machen. Diese Tests sind langsam und sollten als solche gekennzeichnet werden. Hier geht es darum, die Pipeline zu prรคsentieren und sicherzustellen
sicherzustellen, dass es in zukรผnftigen Versionen keine Abweichungen gibt.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/training.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Optimierung eines vortrainierten Modells
[[open-in-colab]]
Die Verwendung eines vorab trainierten Modells hat erhebliche Vorteile. Es reduziert die Rechenkosten und den CO2-Fuรabdruck und ermรถglicht Ihnen die Verwendung von Modellen, die dem neuesten Stand der Technik entsprechen, ohne dass Sie ein Modell von Grund auf neu trainieren mรผssen. Transformers bietet Zugang zu Tausenden von vortrainierten Modellen fรผr eine Vielzahl von Aufgaben. Wenn Sie ein vorab trainiertes Modell verwenden, trainieren Sie es auf einem fรผr Ihre Aufgabe spezifischen Datensatz. Dies wird als Feinabstimmung bezeichnet und ist eine unglaublich leistungsfรคhige Trainingstechnik. In diesem Tutorial werden Sie ein vortrainiertes Modell mit einem Deep-Learning-Framework Ihrer Wahl feinabstimmen:
* Feinabstimmung eines vorab trainierten Modells mit ๐ค Transformers [`Trainer`].
* Feinabstimmung eines vorab trainierten Modells in TensorFlow mit Keras.
* Feinabstimmung eines vorab trainierten Modells in nativem PyTorch.
<a id='data-processing'></a>
## Vorbereitung eines Datensatzes
<Youtube id="_BZearw7f0w"/>
Bevor Sie die Feinabstimmung eines vortrainierten Modells vornehmen kรถnnen, mรผssen Sie einen Datensatz herunterladen und fรผr das Training vorbereiten. Im vorangegangenen Leitfaden haben Sie gelernt, wie man Daten fรผr das Training aufbereitet, und jetzt haben Sie die Gelegenheit, diese Fรคhigkeiten zu testen!
Laden Sie zunรคchst den Datensatz [Yelp Reviews](https://huggingface.co/datasets/yelp_review_full):
```py
>>> from datasets import load_dataset
>>> dataset = load_dataset("yelp_review_full")
>>> dataset["train"][100]
{'label': 0,
'text': 'My expectations for McDonalds are t rarely high. But for one to still fail so spectacularly...that takes something special!\\nThe cashier took my friends\'s order, then promptly ignored me. I had to force myself in front of a cashier who opened his register to wait on the person BEHIND me. I waited over five minutes for a gigantic order that included precisely one kid\'s meal. After watching two people who ordered after me be handed their food, I asked where mine was. The manager started yelling at the cashiers for \\"serving off their orders\\" when they didn\'t have their food. But neither cashier was anywhere near those controls, and the manager was the one serving food to customers and clearing the boards.\\nThe manager was rude when giving me my order. She didn\'t make sure that I had everything ON MY RECEIPT, and never even had the decency to apologize that I felt I was getting poor service.\\nI\'ve eaten at various McDonalds restaurants for over 30 years. I\'ve worked at more than one location. I expect bad days, bad moods, and the occasional mistake. But I have yet to have a decent experience at this store. It will remain a place I avoid unless someone in my party needs to avoid illness from low blood sugar. Perhaps I should go back to the racially biased service of Steak n Shake instead!'}
```
Wie Sie nun wissen, benรถtigen Sie einen Tokenizer, um den Text zu verarbeiten und eine Auffรผll- und Abschneidungsstrategie einzubauen, um mit variablen Sequenzlรคngen umzugehen. Um Ihren Datensatz in einem Schritt zu verarbeiten, verwenden Sie die ๐ค Methode Datasets [`map`](https://huggingface.co/docs/datasets/process#map), um eine Vorverarbeitungsfunktion auf den gesamten Datensatz anzuwenden:
```py
>>> from transformers import AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("bert-base-cased")
>>> def tokenize_function(examples):
... return tokenizer(examples["text"], padding="max_length", truncation=True)
>>> tokenized_datasets = dataset.map(tokenize_function, batched=True)
```
Wenn Sie mรถchten, kรถnnen Sie eine kleinere Teilmenge des gesamten Datensatzes fรผr die Feinabstimmung erstellen, um den Zeitaufwand zu verringern:
```py
>>> small_train_dataset = tokenized_datasets["train"].shuffle(seed=42).select(range(1000))
>>> small_eval_dataset = tokenized_datasets["test"].shuffle(seed=42).select(range(1000))
```
<a id='trainer'></a>
## Training
An dieser Stelle sollten Sie dem Abschnitt folgen, der dem Rahmen entspricht, den Sie verwenden mรถchten. Sie kรถnnen รผber die Links
in der rechten Seitenleiste kรถnnen Sie zu dem gewรผnschten Abschnitt springen - und wenn Sie den gesamten Inhalt eines bestimmten Frameworks ausblenden mรถchten,
klicken Sie einfach auf die Schaltflรคche oben rechts im Block des jeweiligen Frameworks!
<frameworkcontent>
<pt>
<Youtube id="nvBXf7s7vTI"/>
## Trainieren mit PyTorch Trainer
๐ค Transformers bietet eine [`Trainer`]-Klasse, die fรผr das Training von ๐ค Transformers-Modellen optimiert ist und es einfacher macht, mit dem Training zu beginnen, ohne manuell eine eigene Trainingsschleife zu schreiben. Die [`Trainer`]-API unterstรผtzt eine breite Palette von Trainingsoptionen und Funktionen wie Logging, Gradientenakkumulation und gemischte Prรคzision.
Beginnen Sie mit dem Laden Ihres Modells und geben Sie die Anzahl der erwarteten Labels an. Aus dem Yelp Review [dataset card](https://huggingface.co/datasets/yelp_review_full#data-fields) wissen Sie, dass es fรผnf Labels gibt:
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained("bert-base-cased", num_labels=5)
```
<Tip>
Es wird eine Warnung angezeigt, dass einige der trainierten Parameter nicht verwendet werden und einige Parameter zufรคllig
initialisiert werden. Machen Sie sich keine Sorgen, das ist vรถllig normal! Der vorher trainierte Kopf des BERT-Modells wird verworfen und durch einen zufรคllig initialisierten Klassifikationskopf ersetzt. Sie werden diesen neuen Modellkopf in Ihrer Sequenzklassifizierungsaufgabe feinabstimmen, indem Sie das Wissen des vortrainierten Modells auf ihn รผbertragen.
</Tip>
### Hyperparameter fรผr das Training
Als Nรคchstes erstellen Sie eine Klasse [`TrainingArguments`], die alle Hyperparameter enthรคlt, die Sie einstellen kรถnnen, sowie Flags zur Aktivierung verschiedener Trainingsoptionen. Fรผr dieses Lernprogramm kรถnnen Sie mit den Standard- [Hyperparametern](https://huggingface.co/docs/transformers/main_classes/trainer#transformers.TrainingArguments) beginnen, aber Sie kรถnnen mit diesen experimentieren, um Ihre optimalen Einstellungen zu finden.
Geben Sie an, wo die Kontrollpunkte Ihres Trainings gespeichert werden sollen:
```py
>>> from transformers import TrainingArguments
>>> training_args = TrainingArguments(output_dir="test_trainer")
```
### Auswerten
Der [`Trainer`] wertet die Leistung des Modells wรคhrend des Trainings nicht automatisch aus. Sie mรผssen [`Trainer`] eine Funktion รผbergeben, um Metriken zu berechnen und zu berichten. Die [๐ค Evaluate](https://huggingface.co/docs/evaluate/index) Bibliothek bietet eine einfache [`accuracy`](https://huggingface.co/spaces/evaluate-metric/accuracy) Funktion, die Sie mit der [`evaluate.load`] Funktion laden kรถnnen (siehe diese [quicktour](https://huggingface.co/docs/evaluate/a_quick_tour) fรผr weitere Informationen):
```py
>>> import numpy as np
>>> import evaluate
>>> metric = evaluate.load("accuracy")
```
Rufen Sie [`~evaluate.compute`] auf `metric` auf, um die Genauigkeit Ihrer Vorhersagen zu berechnen. Bevor Sie Ihre Vorhersagen an `compute` รผbergeben, mรผssen Sie die Vorhersagen in Logits umwandeln (denken Sie daran, dass alle ๐ค Transformers-Modelle Logits zurรผckgeben):
```py
>>> def compute_metrics(eval_pred):
... logits, labels = eval_pred
... predictions = np.argmax(logits, axis=-1)
... return metric.compute(predictions=predictions, references=labels)
```
Wenn Sie Ihre Bewertungsmetriken wรคhrend der Feinabstimmung รผberwachen mรถchten, geben Sie den Parameter `evaluation_strategy` in Ihren Trainingsargumenten an, um die Bewertungsmetrik am Ende jeder Epoche zu ermitteln:
```py
>>> from transformers import TrainingArguments, Trainer
>>> training_args = TrainingArguments(output_dir="test_trainer", evaluation_strategy="epoch")
```
### Trainer
Erstellen Sie ein [`Trainer`]-Objekt mit Ihrem Modell, Trainingsargumenten, Trainings- und Testdatensรคtzen und einer Evaluierungsfunktion:
```py
>>> trainer = Trainer(
... model=model,
... args=training_args,
... train_dataset=small_train_dataset,
... eval_dataset=small_eval_dataset,
... compute_metrics=compute_metrics,
... )
```
Anschlieรend kรถnnen Sie Ihr Modell durch den Aufruf von [`~transformers.Trainer.train`] optimieren:
```py
>>> trainer.train()
```
</pt>
<tf>
<a id='keras'></a>
<Youtube id="rnTGBy2ax1c"/>
## Trainieren Sie ein TensorFlow-Modell mit Keras
Sie kรถnnen auch ๐ค Transformers Modelle in TensorFlow mit der Keras API trainieren!
### Laden von Daten fรผr Keras
Wenn Sie ein ๐ค Transformers Modell mit der Keras API trainieren wollen, mรผssen Sie Ihren Datensatz in ein Format konvertieren, das
Keras versteht. Wenn Ihr Datensatz klein ist, kรถnnen Sie das Ganze einfach in NumPy-Arrays konvertieren und an Keras รผbergeben.
Probieren wir das zuerst aus, bevor wir etwas Komplizierteres tun.
Laden Sie zunรคchst ein Dataset. Wir werden den CoLA-Datensatz aus dem [GLUE-Benchmark](https://huggingface.co/datasets/glue) verwenden,
da es sich um eine einfache Aufgabe zur Klassifizierung von binรคrem Text handelt, und nehmen vorerst nur den Trainingssplit.
```py
from datasets import load_dataset
dataset = load_dataset("glue", "cola")
dataset = dataset["train"] # Just take the training split for now
```
Als nรคchstes laden Sie einen Tokenizer und tokenisieren die Daten als NumPy-Arrays. Beachten Sie, dass die Beschriftungen bereits eine Liste von 0 und 1en sind,
Wir kรถnnen sie also ohne Tokenisierung direkt in ein NumPy-Array konvertieren!
```py
from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("bert-base-cased")
tokenized_data = tokenizer(dataset["text"], return_tensors="np", padding=True)
# Tokenizer returns a BatchEncoding, but we convert that to a dict for Keras
tokenized_data = dict(tokenized_data)
labels = np.array(dataset["label"]) # Label is already an array of 0 and 1
```
Schlieรlich laden, [`compile`](https://keras.io/api/models/model_training_apis/#compile-method) und [`fit`](https://keras.io/api/models/model_training_apis/#fit-method) Sie das Modell:
```py
from transformers import TFAutoModelForSequenceClassification
from tensorflow.keras.optimizers import Adam
# Load and compile our model
model = TFAutoModelForSequenceClassification.from_pretrained("bert-base-cased")
# Lower learning rates are often better for fine-tuning transformers
model.compile(optimizer=Adam(3e-5))
model.fit(tokenized_data, labels)
```
<Tip>
Sie mรผssen Ihren Modellen kein Verlustargument รผbergeben, wenn Sie sie `compile()`! Hugging-Face-Modelle wรคhlen automatisch
einen Loss, der fรผr ihre Aufgabe und Modellarchitektur geeignet ist, wenn dieses Argument leer gelassen wird. Sie kรถnnen jederzeit auรer Kraft setzen, indem Sie selbst einen Loss angeben, wenn Sie das mรถchten!
</Tip>
Dieser Ansatz eignet sich hervorragend fรผr kleinere Datensรคtze, aber bei grรถรeren Datensรคtzen kann er zu einem Problem werden. Warum?
Weil das tokenisierte Array und die Beschriftungen vollstรคndig in den Speicher geladen werden mรผssten, und weil NumPy nicht mit
"gezackte" Arrays nicht verarbeiten kann, so dass jedes tokenisierte Sample auf die Lรคnge des lรคngsten Samples im gesamten Datensatz aufgefรผllt werden mรผsste.
Datensatzes aufgefรผllt werden. Dadurch wird das Array noch grรถรer, und all die aufgefรผllten Token verlangsamen auch das Training!
### Laden von Daten als tf.data.Dataset
Wenn Sie eine Verlangsamung des Trainings vermeiden wollen, kรถnnen Sie Ihre Daten stattdessen als `tf.data.Dataset` laden. Sie kรถnnen zwar Ihre eigene
tf.data"-Pipeline schreiben kรถnnen, wenn Sie wollen, haben wir zwei bequeme Methoden, um dies zu tun:
- [`~TFPreTrainedModel.prepare_tf_dataset`]: Dies ist die Methode, die wir in den meisten Fรคllen empfehlen. Da es sich um eine Methode
Ihres Modells ist, kann sie das Modell inspizieren, um automatisch herauszufinden, welche Spalten als Modelleingaben verwendet werden kรถnnen, und
verwirft die anderen, um einen einfacheren, leistungsfรคhigeren Datensatz zu erstellen.
- [~datasets.Dataset.to_tf_dataset`]: Diese Methode ist eher auf niedriger Ebene angesiedelt und ist nรผtzlich, wenn Sie genau kontrollieren wollen, wie
Dataset erstellt wird, indem man genau angibt, welche `columns` und `label_cols` einbezogen werden sollen.
Bevor Sie [~TFPreTrainedModel.prepare_tf_dataset`] verwenden kรถnnen, mรผssen Sie die Tokenizer-Ausgaben als Spalten zu Ihrem Datensatz hinzufรผgen, wie in
dem folgenden Codebeispiel:
```py
def tokenize_dataset(data):
# Keys of the returned dictionary will be added to the dataset as columns
return tokenizer(data["text"])
dataset = dataset.map(tokenize_dataset)
```
Denken Sie daran, dass Hugging Face-Datensรคtze standardmรครig auf der Festplatte gespeichert werden, so dass dies nicht zu einem erhรถhten Arbeitsspeicherbedarf fรผhren wird! Sobald die
Spalten hinzugefรผgt wurden, kรถnnen Sie Batches aus dem Datensatz streamen und zu jedem Batch Auffรผllungen hinzufรผgen, was die Anzahl der Auffรผllungs-Token im Vergleich zum Auffรผllen des gesamten Datensatzes reduziert.
```py
>>> tf_dataset = model.prepare_tf_dataset(dataset, batch_size=16, shuffle=True, tokenizer=tokenizer)
```
Beachten Sie, dass Sie im obigen Codebeispiel den Tokenizer an `prepare_tf_dataset` รผbergeben mรผssen, damit die Stapel beim Laden korrekt aufgefรผllt werden kรถnnen.
Wenn alle Stichproben in Ihrem Datensatz die gleiche Lรคnge haben und kein Auffรผllen erforderlich ist, kรถnnen Sie dieses Argument weglassen.
Wenn Sie etwas Komplexeres als nur das Auffรผllen von Stichproben benรถtigen (z. B. das Korrumpieren von Token fรผr die maskierte Sprachmodellierung), kรถnnen Sie das Argument
Modellierung), kรถnnen Sie stattdessen das Argument `collate_fn` verwenden, um eine Funktion zu รผbergeben, die aufgerufen wird, um die
Liste von Stichproben in einen Stapel umwandelt und alle gewรผnschten Vorverarbeitungen vornimmt. Siehe unsere
[examples](https://github.com/huggingface/transformers/tree/main/examples) oder
[notebooks](https://huggingface.co/docs/transformers/notebooks), um diesen Ansatz in Aktion zu sehen.
Sobald Sie einen `tf.data.Dataset` erstellt haben, kรถnnen Sie das Modell wie zuvor kompilieren und anpassen:
```py
model.compile(optimizer=Adam(3e-5))
model.fit(tf_dataset)
```
</tf>
</frameworkcontent>
<a id='pytorch_native'></a>
## Trainieren in nativem PyTorch
<frameworkcontent>
<pt>
<Youtube id="Dh9CL8fyG80"/>
[`Trainer`] kรผmmert sich um die Trainingsschleife und ermรถglicht die Feinabstimmung eines Modells in einer einzigen Codezeile. Fรผr Benutzer, die es vorziehen, ihre eigene Trainingsschleife zu schreiben, kรถnnen Sie auch eine Feinabstimmung eines ๐ค Transformers-Modells in nativem PyTorch vornehmen.
An diesem Punkt mรผssen Sie mรถglicherweise Ihr Notebook neu starten oder den folgenden Code ausfรผhren, um etwas Speicher freizugeben:
```py
del model
del pytorch_model
del trainer
torch.cuda.empty_cache()
```
Als Nรคchstes mรผssen Sie den Datensatz `tokenized_dataset` manuell nachbearbeiten, um ihn fรผr das Training vorzubereiten.
1. Entfernen Sie die Spalte "Text", da das Modell keinen Rohtext als Eingabe akzeptiert:
```py
>>> tokenized_datasets = tokenized_datasets.remove_columns(["text"])
```
2. Benennen Sie die Spalte "Label" in "Labels" um, da das Modell erwartet, dass das Argument "Labels" genannt wird:
```py
>>> tokenized_datasets = tokenized_datasets.rename_column("label", "labels")
```
3. Stellen Sie das Format des Datensatzes so ein, dass PyTorch-Tensoren anstelle von Listen zurรผckgegeben werden:
```py
>>> tokenized_datasets.set_format("torch")
```
Erstellen Sie dann eine kleinere Teilmenge des Datensatzes, wie zuvor gezeigt, um die Feinabstimmung zu beschleunigen:
```py
>>> small_train_dataset = tokenized_datasets["train"].shuffle(seed=42).select(range(1000))
>>> small_eval_dataset = tokenized_datasets["test"].shuffle(seed=42).select(range(1000))
```
### DataLoader
Erstellen Sie einen `DataLoader` fรผr Ihre Trainings- und Testdatensรคtze, damit Sie รผber die Datenstapel iterieren kรถnnen:
```py
>>> from torch.utils.data import DataLoader
>>> train_dataloader = DataLoader(small_train_dataset, shuffle=True, batch_size=8)
>>> eval_dataloader = DataLoader(small_eval_dataset, batch_size=8)
```
Laden Sie Ihr Modell mit der Anzahl der erwarteten Kennzeichnungen:
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained("bert-base-cased", num_labels=5)
```
### Optimierer und Lernratensteuerung
Erstellen Sie einen Optimierer und einen Scheduler fรผr die Lernrate, um das Modell fein abzustimmen. Wir verwenden den Optimierer [`AdamW`](https://pytorch.org/docs/stable/generated/torch.optim.AdamW.html) aus PyTorch:
```py
>>> from torch.optim import AdamW
>>> optimizer = AdamW(model.parameters(), lr=5e-5)
```
Erstellen Sie den Standard-Lernratenplaner aus [`Trainer`]:
```py
>>> from transformers import get_scheduler
>>> num_epochs = 3
>>> num_training_steps = num_epochs * len(train_dataloader)
>>> lr_scheduler = get_scheduler(
... name="linear", optimizer=optimizer, num_warmup_steps=0, num_training_steps=num_training_steps
... )
```
Geben Sie schlieรlich `device` an, um einen Grafikprozessor zu verwenden, wenn Sie Zugang zu einem solchen haben. Andernfalls kann das Training auf einer CPU mehrere Stunden statt ein paar Minuten dauern.
```py
>>> import torch
>>> device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
>>> model.to(device)
```
<Tip>
Holen Sie sich mit einem gehosteten Notebook wie [Colaboratory](https://colab.research.google.com/) oder [SageMaker StudioLab](https://studiolab.sagemaker.aws/) kostenlosen Zugang zu einem Cloud-GPU, wenn Sie noch keinen haben.
</Tip>
Groรartig, Sie sind bereit fรผr das Training! ๐ฅณ
### Trainingsschleife
Um Ihren Trainingsfortschritt zu verfolgen, verwenden Sie die [tqdm](https://tqdm.github.io/) Bibliothek, um einen Fortschrittsbalken รผber die Anzahl der Trainingsschritte hinzuzufรผgen:
```py
>>> from tqdm.auto import tqdm
>>> progress_bar = tqdm(range(num_training_steps))
>>> model.train()
>>> for epoch in range(num_epochs):
... for batch in train_dataloader:
... batch = {k: v.to(device) for k, v in batch.items()}
... outputs = model(**batch)
... loss = outputs.loss
... loss.backward()
... optimizer.step()
... lr_scheduler.step()
... optimizer.zero_grad()
... progress_bar.update(1)
```
### Auswertung
Genauso wie Sie eine Bewertungsfunktion zu [`Trainer`] hinzugefรผgt haben, mรผssen Sie dasselbe tun, wenn Sie Ihre eigene Trainingsschleife schreiben. Aber anstatt die Metrik am Ende jeder Epoche zu berechnen und zu melden, werden Sie dieses Mal alle Stapel mit [`~evaluate.add_batch`] akkumulieren und die Metrik ganz am Ende berechnen.
```py
>>> import evaluate
>>> metric = evaluate.load("accuracy")
>>> model.eval()
>>> for batch in eval_dataloader:
... batch = {k: v.to(device) for k, v in batch.items()}
... with torch.no_grad():
... outputs = model(**batch)
... logits = outputs.logits
... predictions = torch.argmax(logits, dim=-1)
... metric.add_batch(predictions=predictions, references=batch["labels"])
>>> metric.compute()
```
</pt>
</frameworkcontent>
<a id='additional-resources'></a>
## Zusรคtzliche Ressourcen
Weitere Beispiele fรผr die Feinabstimmung finden Sie unter:
- [๐ค Transformers Examples](https://github.com/huggingface/transformers/tree/main/examples) enthรคlt Skripte
um gรคngige NLP-Aufgaben in PyTorch und TensorFlow zu trainieren.
- [๐ค Transformers Notebooks](notebooks) enthรคlt verschiedene Notebooks zur Feinabstimmung eines Modells fรผr bestimmte Aufgaben in PyTorch und TensorFlow.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/index.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ๐ค Transformers
Maschinelles Lernen auf dem neuesten Stand der Technik fรผr PyTorch, TensorFlow und JAX.
๐ค Transformers bietet APIs zum einfachen Herunterladen und Trainieren von vortrainierten Modellen auf dem neuesten Stand der Technik. Die Verwendung von vortrainierten Modellen kann Rechenkosten sparen und den CO2-Fuรabdruck reduzieren und Zeit sparen, die fรผr das Training eines Modells von Grund auf benรถtigt wird. Die Modelle kรถnnen fรผr verschiedene Modalitรคten verwendet werden, wie z. B.:
* ๐ Text: Textklassifizierung, Informationsextrahierung, Beantwortung von Fragen, Zusammenfassung, รbersetzung und Texterstellung in รผber 100 Sprachen.
* ๐ผ๏ธ Bilder: Bildklassifizierung, Objekterkennung und Segmentierung.
* ๐ฃ๏ธ Audio: Spracherkennung und Audioklassifizierung.
* ๐ Multimodal: Beantwortung von Tabellenfragen, optische Zeichenerkennung, Informationsextraktion aus gescannten Dokumenten, Videoklassifizierung und Beantwortung visueller Fragen.
Unsere Bibliothek unterstรผtzt die nahtlose Integration von drei der beliebtesten Deep-Learning-Bibliotheken: [PyTorch](https://pytorch.org/), [TensorFlow](https://www.tensorflow.org/) und [JAX](https://jax.readthedocs.io/en/latest/). Trainieren Sie Ihr Modell in drei Codezeilen in einem Framework und laden Sie es zur Inferenz mit einem anderen.
Jede ๐ค Transformers-Architektur ist in einem eigenstรคndigen Python-Modul definiert, so dass sie leicht fรผr Forschung und Experimente angepasst werden kann.
## Wenn Sie auf der Suche nach individueller Unterstรผtzung durch das Hugging Face-Team sind
<a target="_blank" href="https://huggingface.co/support">
<img alt="HuggingFace Expert Acceleration Program" src="https://cdn-media.huggingface.co/marketing/transformers/new-support-improved.png" style="width: 100%; max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
</a>
## Inhalt
Die Dokumentation ist in fรผnf Teile gegliedert:
- **GET STARTED** enthรคlt eine kurze Tour und Installationsanweisungen, um mit ๐ค Transformers loszulegen.
- **TUTORIALS** sind ein hervorragender Ausgangspunkt, wenn Sie neu in unserer Bibliothek sind. Dieser Abschnitt hilft Ihnen, die grundlegenden Fรคhigkeiten zu erlangen, die Sie benรถtigen, um mit ๐ค Transformers zu arbeiten.
- **HOW-TO GUIDES** zeigen Ihnen, wie Sie ein bestimmtes Ziel erreichen kรถnnen, z. B. die Feinabstimmung eines vortrainierten Modells fรผr die Sprachmodellierung oder die Erstellung eines benutzerdefinierten Modellkopfs.
- **KONZEPTUELLE ANLEITUNGEN** bietet weitere Diskussionen und Erklรคrungen zu den zugrunde liegenden Konzepten und Ideen hinter Modellen, Aufgaben und der Designphilosophie von ๐ค Transformers.
- **API** beschreibt jede Klasse und Funktion, gruppiert in:
- **MAIN CLASSES** fรผr die Hauptklassen, die die wichtigsten APIs der Bibliothek darstellen.
- MODELLE** fรผr die Klassen und Funktionen, die zu jedem in der Bibliothek implementierten Modell gehรถren.
- **INTERNAL HELPERS** fรผr die Klassen und Funktionen, die wir intern verwenden.
Die Bibliothek enthรคlt derzeit JAX-, PyTorch- und TensorFlow-Implementierungen, vortrainierte Modellgewichte, Nutzungsskripte und Konvertierungsprogramme fรผr die folgenden Modelle.
### Unterstรผtze Modelle
<!--This list is updated automatically from the README with _make fix-copies_. Do not update manually! -->
1. **[ALBERT](model_doc/albert)** (from Google Research and the Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
1. **[ALIGN](model_doc/align)** (from Google Research) released with the paper [Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision](https://arxiv.org/abs/2102.05918) by Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc V. Le, Yunhsuan Sung, Zhen Li, Tom Duerig.
1. **[BART](model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
1. **[BARThez](model_doc/barthez)** (from รcole polytechnique) released with the paper [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) by Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis.
1. **[BARTpho](model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
1. **[BEiT](model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
1. **[BERT](model_doc/bert)** (from Google) released with the paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.
1. **[BERT For Sequence Generation](model_doc/bert-generation)** (from Google) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
1. **[BERTweet](model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
1. **[BigBird-Pegasus](model_doc/bigbird_pegasus)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
1. **[BigBird-RoBERTa](model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
1. **[Blenderbot](model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
1. **[BlenderbotSmall](model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
1. **[BLOOM](model_doc/bloom)** (from BigScience workshop) released by the [BigScience Workshop](https://bigscience.huggingface.co/).
1. **[BORT](model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
1. **[ByT5](model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
1. **[CamemBERT](model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suรกrez*, Yoann Dupont, Laurent Romary, รric Villemonte de la Clergerie, Djamรฉ Seddah and Benoรฎt Sagot.
1. **[CANINE](model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
1. **[CLIP](model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
1. **[CodeGen](model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
1. **[ConvBERT](model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
1. **[ConvNeXT](model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
1. **[ConvNeXTV2](model_doc/convnextv2)** (from Facebook AI) released with the paper [ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders](https://arxiv.org/abs/2301.00808) by Sanghyun Woo, Shoubhik Debnath, Ronghang Hu, Xinlei Chen, Zhuang Liu, In So Kweon, Saining Xie.
1. **[CPM](model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
1. **[CTRL](model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
1. **[CvT](model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
1. **[Data2Vec](model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec: A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
1. **[DeBERTa](model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
1. **[DeBERTa-v2](model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
1. **[Decision Transformer](model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
1. **[DeiT](model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervรฉ Jรฉgou.
1. **[DETR](model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
1. **[DialoGPT](model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
1. **[DistilBERT](model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation) and a German version of DistilBERT.
1. **[DiT](model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
1. **[DPR](model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oฤuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
1. **[DPT](master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by Renรฉ Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
1. **[EfficientNet](model_doc/efficientnet)** (from Google Research) released with the paper [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://arxiv.org/abs/1905.11946) by Mingxing Tan and Quoc V. Le.
1. **[ELECTRA](model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
1. **[EncoderDecoder](model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
1. **[FlauBERT](model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loรฏc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoรฎt Crabbรฉ, Laurent Besacier, Didier Schwab.
1. **[FLAVA](model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
1. **[FNet](model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
1. **[Funnel Transformer](model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
1. **[GLPN](model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
1. **[GPT](model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
1. **[GPT Neo](model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
1. **[GPT NeoX](model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach
1. **[GPT-2](model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
1. **[GPT-J](model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
1. **[GPTSAN-japanese](model_doc/gptsan-japanese)** released in the repository [tanreinama/GPTSAN](https://github.com/tanreinama/GPTSAN/blob/main/report/model.md) by Toshiyuki Sakamoto(tanreinama).
1. **[GroupViT](model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
1. **[Hubert](model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
1. **[I-BERT](model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
1. **[ImageGPT](model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
1. **[LayoutLM](model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
1. **[LayoutLMv2](model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
1. **[LayoutLMv3](model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
1. **[LayoutXLM](model_doc/layoutxlm)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
1. **[LED](model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
1. **[LeViT](model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervรฉ Jรฉgou, Matthijs Douze.
1. **[Longformer](model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
1. **[LongT5](model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
1. **[LUKE](model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
1. **[LXMERT](model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
1. **[M-CTC-T](model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
1. **[M2M100](model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
1. **[MarianMT](model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jรถrg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
1. **[Mask2Former](model_doc/mask2former)** (from FAIR and UIUC) released with the paper [Masked-attention Mask Transformer for Universal Image Segmentation](https://arxiv.org/abs/2112.01527) by Bowen Cheng, Ishan Misra, Alexander G. Schwing, Alexander Kirillov, Rohit Girdhar.
1. **[MaskFormer](model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
1. **[mBART](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
1. **[mBART-50](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
1. **[Megatron-BERT](model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
1. **[Megatron-GPT2](model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
1. **[mLUKE](model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
1. **[MobileBERT](model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
1. **[MobileViT](model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
1. **[MPNet](model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
1. **[MT5](model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
1. **[MVP](model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
1. **[Nezha](model_doc/nezha)** (from Huawei Noahโs Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
1. **[NLLB](model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
1. **[Nystrรถmformer](model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nystrรถmformer: A Nystrรถm-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
1. **[OneFormer](model_doc/oneformer)** (from SHI Labs) released with the paper [OneFormer: One Transformer to Rule Universal Image Segmentation](https://arxiv.org/abs/2211.06220) by Jitesh Jain, Jiachen Li, MangTik Chiu, Ali Hassani, Nikita Orlov, Humphrey Shi.
1. **[OPT](master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
1. **[OWL-ViT](model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
1. **[Pegasus](model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
1. **[Perceiver IO](model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hรฉnaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, Joรฃo Carreira.
1. **[PhoBERT](model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
1. **[PLBart](model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
1. **[PoolFormer](model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
1. **[ProphetNet](model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
1. **[QDQBert](model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
1. **[RAG](model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Kรผttler, Mike Lewis, Wen-tau Yih, Tim Rocktรคschel, Sebastian Riedel, Douwe Kiela.
1. **[REALM](model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
1. **[Reformer](model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, ลukasz Kaiser, Anselm Levskaya.
1. **[RegNet](model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollรกr.
1. **[RemBERT](model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Fรฉvry, Henry Tsai, M. Johnson, Sebastian Ruder.
1. **[ResNet](model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.
1. **[RoBERTa](model_doc/roberta)** (from Facebook), released together with the paper [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
1. **[RoFormer](model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
1. **[SegFormer](model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
1. **[SEW](model_doc/sew)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
1. **[SEW-D](model_doc/sew_d)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
1. **[SpeechToTextTransformer](model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
1. **[SpeechToTextTransformer2](model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
1. **[Splinter](model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
1. **[SqueezeBERT](model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
1. **[Swin Transformer](model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
1. **[Swin Transformer V2](model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
1. **[T5](model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
1. **[T5v1.1](model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
1. **[TAPAS](model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweล Krzysztof Nowak, Thomas Mรผller, Francesco Piccinno and Julian Martin Eisenschlos.
1. **[TAPEX](model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
1. **[Trajectory Transformer](model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine
1. **[Transformer-XL](model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
1. **[TrOCR](model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
1. **[UL2](model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler
1. **[UMT5](model_doc/umt5)** (from Google Research) released with the paper [UniMax: Fairer and More Effective Language Sampling for Large-Scale Multilingual Pretraining](https://openreview.net/forum?id=kXwdL1cWOAi) by Hyung Won Chung, Xavier Garcia, Adam Roberts, Yi Tay, Orhan Firat, Sharan Narang, Noah Constant.
1. **[UniSpeech](model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
1. **[UniSpeechSat](model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
1. **[VAN](model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/abs/2202.09741) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
1. **[VideoMAE](model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
1. **[ViLT](model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
1. **[Vision Transformer (ViT)](model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
1. **[VisualBERT](model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
1. **[ViTMAE](model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollรกr, Ross Girshick.
1. **[Wav2Vec2](model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
1. **[Wav2Vec2-Conformer](model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
1. **[Wav2Vec2Phoneme](model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
1. **[WavLM](model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
1. **[XGLM](model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li.
1. **[XLM](model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
1. **[XLM-ProphetNet](model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
1. **[XLM-RoBERTa](model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmรกn, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
1. **[XLM-RoBERTa-XL](model_doc/xlm-roberta-xl)** (from Facebook AI), released together with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
1. **[XLM-V](model_doc/xlm-v)** (from Meta AI) released with the paper [XLM-V: Overcoming the Vocabulary Bottleneck in Multilingual Masked Language Models](https://arxiv.org/abs/2301.10472) by Davis Liang, Hila Gonen, Yuning Mao, Rui Hou, Naman Goyal, Marjan Ghazvininejad, Luke Zettlemoyer, Madian Khabsa.
1. **[XLNet](model_doc/xlnet)** (from Google/CMU) released with the paper [โXLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
1. **[XLS-R](model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
1. **[XLSR-Wav2Vec2](model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
1. **[YOLOS](model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
1. **[YOSO](model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
### Unterstรผtzte Frameworks
Die folgende Tabelle zeigt die derzeitige Unterstรผtzung in der Bibliothek fรผr jedes dieser Modelle, unabhรคngig davon, ob sie einen Python
Tokenizer haben (als "langsam" bezeichnet), ein "schneller" Tokenizer, der von der ๐ค Tokenizers Bibliothek unterstรผtzt wird, ob sie Unterstรผtzung in Jax (via
Flax), PyTorch, und/oder TensorFlow haben.
<!--This table is updated automatically from the auto modules with _make fix-copies_. Do not update manually!-->
| Model | Tokenizer slow | Tokenizer fast | PyTorch support | TensorFlow support | Flax Support |
|:---------------------------:|:--------------:|:--------------:|:---------------:|:------------------:|:------------:|
| ALBERT | โ
| โ
| โ
| โ
| โ
|
| BART | โ
| โ
| โ
| โ
| โ
|
| BEiT | โ | โ | โ
| โ | โ
|
| BERT | โ
| โ
| โ
| โ
| โ
|
| Bert Generation | โ
| โ | โ
| โ | โ |
| BigBird | โ
| โ
| โ
| โ | โ
|
| BigBird-Pegasus | โ | โ | โ
| โ | โ |
| Blenderbot | โ
| โ
| โ
| โ
| โ
|
| BlenderbotSmall | โ
| โ
| โ
| โ
| โ
|
| BLOOM | โ | โ
| โ
| โ | โ
|
| CamemBERT | โ
| โ
| โ
| โ
| โ |
| CANINE | โ
| โ | โ
| โ | โ |
| CLIP | โ
| โ
| โ
| โ
| โ
|
| CodeGen | โ
| โ
| โ
| โ | โ |
| ConvBERT | โ
| โ
| โ
| โ
| โ |
| ConvNeXT | โ | โ | โ
| โ
| โ |
| CTRL | โ
| โ | โ
| โ
| โ |
| CvT | โ | โ | โ
| โ | โ |
| Data2VecAudio | โ | โ | โ
| โ | โ |
| Data2VecText | โ | โ | โ
| โ | โ |
| Data2VecVision | โ | โ | โ
| โ
| โ |
| DeBERTa | โ
| โ
| โ
| โ
| โ |
| DeBERTa-v2 | โ
| โ
| โ
| โ
| โ |
| Decision Transformer | โ | โ | โ
| โ | โ |
| DeiT | โ | โ | โ
| โ
| โ |
| DETR | โ | โ | โ
| โ | โ |
| DistilBERT | โ
| โ
| โ
| โ
| โ
|
| DPR | โ
| โ
| โ
| โ
| โ |
| DPT | โ | โ | โ
| โ | โ |
| ELECTRA | โ
| โ
| โ
| โ
| โ
|
| Encoder decoder | โ | โ | โ
| โ
| โ
|
| FairSeq Machine-Translation | โ
| โ | โ
| โ | โ |
| FlauBERT | โ
| โ | โ
| โ
| โ |
| FLAVA | โ | โ | โ
| โ | โ |
| FNet | โ
| โ
| โ
| โ | โ |
| Funnel Transformer | โ
| โ
| โ
| โ
| โ |
| GLPN | โ | โ | โ
| โ | โ |
| GPT Neo | โ | โ | โ
| โ | โ
|
| GPT NeoX | โ | โ
| โ
| โ | โ |
| GPT-J | โ | โ | โ
| โ
| โ
|
| GroupViT | โ | โ | โ
| โ | โ |
| Hubert | โ | โ | โ
| โ
| โ |
| I-BERT | โ | โ | โ
| โ | โ |
| ImageGPT | โ | โ | โ
| โ | โ |
| LayoutLM | โ
| โ
| โ
| โ
| โ |
| LayoutLMv2 | โ
| โ
| โ
| โ | โ |
| LayoutLMv3 | โ
| โ
| โ
| โ | โ |
| LED | โ
| โ
| โ
| โ
| โ |
| LeViT | โ | โ | โ
| โ | โ |
| Longformer | โ
| โ
| โ
| โ
| โ |
| LongT5 | โ | โ | โ
| โ | โ
|
| LUKE | โ
| โ | โ
| โ | โ |
| LXMERT | โ
| โ
| โ
| โ
| โ |
| M-CTC-T | โ | โ | โ
| โ | โ |
| M2M100 | โ
| โ | โ
| โ | โ |
| Marian | โ
| โ | โ
| โ
| โ
|
| MaskFormer | โ | โ | โ
| โ | โ |
| mBART | โ
| โ
| โ
| โ
| โ
|
| Megatron-BERT | โ | โ | โ
| โ | โ |
| MobileBERT | โ
| โ
| โ
| โ
| โ |
| MobileViT | โ | โ | โ
| โ | โ |
| MPNet | โ
| โ
| โ
| โ
| โ |
| MT5 | โ
| โ
| โ
| โ
| โ
|
| MVP | โ
| โ
| โ
| โ | โ |
| Nezha | โ | โ | โ
| โ | โ |
| Nystrรถmformer | โ | โ | โ
| โ | โ |
| OpenAI GPT | โ
| โ
| โ
| โ
| โ |
| OpenAI GPT-2 | โ
| โ
| โ
| โ
| โ
|
| OPT | โ | โ | โ
| โ
| โ
|
| OWL-ViT | โ | โ | โ
| โ | โ |
| Pegasus | โ
| โ
| โ
| โ
| โ
|
| Perceiver | โ
| โ | โ
| โ | โ |
| PLBart | โ
| โ | โ
| โ | โ |
| PoolFormer | โ | โ | โ
| โ | โ |
| ProphetNet | โ
| โ | โ
| โ | โ |
| QDQBert | โ | โ | โ
| โ | โ |
| RAG | โ
| โ | โ
| โ
| โ |
| REALM | โ
| โ
| โ
| โ | โ |
| Reformer | โ
| โ
| โ
| โ | โ |
| RegNet | โ | โ | โ
| โ
| โ
|
| RemBERT | โ
| โ
| โ
| โ
| โ |
| ResNet | โ | โ | โ
| โ
| โ
|
| RetriBERT | โ
| โ
| โ
| โ | โ |
| RoBERTa | โ
| โ
| โ
| โ
| โ
|
| RoFormer | โ
| โ
| โ
| โ
| โ
|
| SegFormer | โ | โ | โ
| โ
| โ |
| SEW | โ | โ | โ
| โ | โ |
| SEW-D | โ | โ | โ
| โ | โ |
| Speech Encoder decoder | โ | โ | โ
| โ | โ
|
| Speech2Text | โ
| โ | โ
| โ
| โ |
| Speech2Text2 | โ
| โ | โ | โ | โ |
| Splinter | โ
| โ
| โ
| โ | โ |
| SqueezeBERT | โ
| โ
| โ
| โ | โ |
| Swin Transformer | โ | โ | โ
| โ
| โ |
| Swin Transformer V2 | โ | โ | โ
| โ | โ |
| T5 | โ
| โ
| โ
| โ
| โ
|
| TAPAS | โ
| โ | โ
| โ
| โ |
| Trajectory Transformer | โ | โ | โ
| โ | โ |
| Transformer-XL | โ
| โ | โ
| โ
| โ |
| TrOCR | โ | โ | โ
| โ | โ |
| UniSpeech | โ | โ | โ
| โ | โ |
| UniSpeechSat | โ | โ | โ
| โ | โ |
| VAN | โ | โ | โ
| โ | โ |
| VideoMAE | โ | โ | โ
| โ | โ |
| ViLT | โ | โ | โ
| โ | โ |
| Vision Encoder decoder | โ | โ | โ
| โ
| โ
|
| VisionTextDualEncoder | โ | โ | โ
| โ | โ
|
| VisualBERT | โ | โ | โ
| โ | โ |
| ViT | โ | โ | โ
| โ
| โ
|
| ViTMAE | โ | โ | โ
| โ
| โ |
| Wav2Vec2 | โ
| โ | โ
| โ
| โ
|
| Wav2Vec2-Conformer | โ | โ | โ
| โ | โ |
| WavLM | โ | โ | โ
| โ | โ |
| XGLM | โ
| โ
| โ
| โ | โ
|
| XLM | โ
| โ | โ
| โ
| โ |
| XLM-ProphetNet | โ
| โ | โ
| โ | โ |
| XLM-RoBERTa | โ
| โ
| โ
| โ
| โ
|
| XLM-RoBERTa-XL | โ | โ | โ
| โ | โ |
| XLNet | โ
| โ
| โ
| โ
| โ |
| YOLOS | โ | โ | โ
| โ | โ |
| YOSO | โ | โ | โ
| โ | โ |
<!-- End table-->
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/llm_tutorial.md
|
<!--Copyright 2023 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Generation with LLMs
[[open-in-colab]]
LLMs (Large Language Models) sind die Schlรผsselkomponente bei der Texterstellung. Kurz gesagt, bestehen sie aus groรen, vortrainierten Transformationsmodellen, die darauf trainiert sind, das nรคchste Wort (oder genauer gesagt Token) aus einem Eingabetext vorherzusagen. Da sie jeweils ein Token vorhersagen, mรผssen Sie etwas Aufwรคndigeres tun, um neue Sรคtze zu generieren, als nur das Modell aufzurufen - Sie mรผssen eine autoregressive Generierung durchfรผhren.
Die autoregressive Generierung ist ein Verfahren zur Inferenzzeit, bei dem ein Modell mit seinen eigenen generierten Ausgaben iterativ aufgerufen wird, wenn einige anfรคngliche Eingaben vorliegen. In ๐ค Transformers wird dies von der Methode [`~generation.GenerationMixin.generate`] รผbernommen, die allen Modellen mit generativen Fรคhigkeiten zur Verfรผgung steht.
Dieses Tutorial zeigt Ihnen, wie Sie:
* Text mit einem LLM generieren
* Vermeiden Sie hรคufige Fallstricke
* Nรคchste Schritte, damit Sie das Beste aus Ihrem LLM herausholen kรถnnen
Bevor Sie beginnen, stellen Sie sicher, dass Sie alle erforderlichen Bibliotheken installiert haben:
```bash
pip install transformers bitsandbytes>=0.39.0 -q
```
## Text generieren
Ein Sprachmodell, das fรผr [causal language modeling](tasks/language_modeling) trainiert wurde, nimmt eine Folge von Text-Token als Eingabe und gibt die Wahrscheinlichkeitsverteilung fรผr das nรคchste Token zurรผck.
<!-- [GIF 1 -- FWD PASS] -->
<figure class="image table text-center m-0 w-full">
<video
style="max-width: 90%; margin: auto;"
autoplay loop muted playsinline
src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/blog/assisted-generation/gif_1_1080p.mov"
></video>
<figcaption>"Forward pass of an LLM"</figcaption>
</figure>
Ein wichtiger Aspekt der autoregressiven Generierung mit LLMs ist die Auswahl des nรคchsten Tokens aus dieser Wahrscheinlichkeitsverteilung. In diesem Schritt ist alles mรถglich, solange Sie am Ende ein Token fรผr die nรคchste Iteration haben. Das heiรt, es kann so einfach sein wie die Auswahl des wahrscheinlichsten Tokens aus der Wahrscheinlichkeitsverteilung oder so komplex wie die Anwendung von einem Dutzend Transformationen vor der Stichprobenziehung aus der resultierenden Verteilung.
<!-- [GIF 2 -- TEXT GENERATION] -->
<figure class="image table text-center m-0 w-full">
<video
style="max-width: 90%; margin: auto;"
autoplay loop muted playsinline
src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/blog/assisted-generation/gif_2_1080p.mov"
></video>
<figcaption>"Die autoregressive Generierung wรคhlt iterativ das nรคchste Token aus einer Wahrscheinlichkeitsverteilung aus, um Text zu erzeugen"</figcaption>
</figure>
Der oben dargestellte Prozess wird iterativ wiederholt, bis eine bestimmte Abbruchbedingung erreicht ist. Im Idealfall wird die Abbruchbedingung vom Modell vorgegeben, das lernen sollte, wann es ein Ende-der-Sequenz-Token (EOS) ausgeben muss. Ist dies nicht der Fall, stoppt die Generierung, wenn eine vordefinierte Maximallรคnge erreicht ist.
Damit sich Ihr Modell so verhรคlt, wie Sie es fรผr Ihre Aufgabe erwarten, mรผssen Sie den Schritt der Token-Auswahl und die Abbruchbedingung richtig einstellen. Aus diesem Grund haben wir zu jedem Modell eine [`~generation.GenerationConfig`]-Datei, die eine gute generative Standardparametrisierung enthรคlt und zusammen mit Ihrem Modell geladen wird.
Lassen Sie uns รผber Code sprechen!
<Tip>
Wenn Sie an der grundlegenden Verwendung von LLMs interessiert sind, ist unsere High-Level-Schnittstelle [`Pipeline`](pipeline_tutorial) ein guter Ausgangspunkt. LLMs erfordern jedoch oft fortgeschrittene Funktionen wie Quantisierung und Feinsteuerung des Token-Auswahlschritts, was am besten รผber [`~generation.GenerationMixin.generate`] erfolgt. Die autoregressive Generierung mit LLMs ist ebenfalls ressourcenintensiv und sollte fรผr einen angemessenen Durchsatz auf einer GPU ausgefรผhrt werden.
</Tip>
<!-- TODO: update example to llama 2 (or a newer popular baseline) when it becomes ungated -->
Zunรคchst mรผssen Sie das Modell laden.
```py
>>> from transformers import AutoModelForCausalLM
>>> model = AutoModelForCausalLM.from_pretrained(
... "openlm-research/open_llama_7b", device_map="auto", load_in_4bit=True
... )
```
Sie werden zwei Flags in dem Aufruf `from_pretrained` bemerken:
- `device_map` stellt sicher, dass das Modell auf Ihre GPU(s) รผbertragen wird
- `load_in_4bit` wendet [dynamische 4-Bit-Quantisierung](main_classes/quantization) an, um die Ressourcenanforderungen massiv zu reduzieren
Es gibt noch andere Mรถglichkeiten, ein Modell zu initialisieren, aber dies ist eine gute Grundlage, um mit einem LLM zu beginnen.
Als nรคchstes mรผssen Sie Ihre Texteingabe mit einem [tokenizer](tokenizer_summary) vorverarbeiten.
```py
>>> from transformers import AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("openlm-research/open_llama_7b")
>>> model_inputs = tokenizer(["A list of colors: red, blue"], return_tensors="pt").to("cuda")
```
Die Variable `model_inputs` enthรคlt die tokenisierte Texteingabe sowie die Aufmerksamkeitsmaske. Obwohl [`~generation.GenerationMixin.generate`] sein Bestes tut, um die Aufmerksamkeitsmaske abzuleiten, wenn sie nicht รผbergeben wird, empfehlen wir, sie fรผr optimale Ergebnisse wann immer mรถglich zu รผbergeben.
Rufen Sie schlieรlich die Methode [~generation.GenerationMixin.generate] auf, um die generierten Token zurรผckzugeben, die vor dem Drucken in Text umgewandelt werden sollten.
```py
>>> generated_ids = model.generate(**model_inputs)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
'A list of colors: red, blue, green, yellow, black, white, and brown'
```
Und das war's! Mit ein paar Zeilen Code kรถnnen Sie sich die Macht eines LLM zunutze machen.
## Hรคufige Fallstricke
Es gibt viele [Generierungsstrategien](generation_strategies), und manchmal sind die Standardwerte fรผr Ihren Anwendungsfall vielleicht nicht geeignet. Wenn Ihre Ausgaben nicht mit dem รผbereinstimmen, was Sie erwarten, haben wir eine Liste der hรคufigsten Fallstricke erstellt und wie Sie diese vermeiden kรถnnen.
```py
>>> from transformers import AutoModelForCausalLM, AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("openlm-research/open_llama_7b")
>>> tokenizer.pad_token = tokenizer.eos_token # Llama has no pad token by default
>>> model = AutoModelForCausalLM.from_pretrained(
... "openlm-research/open_llama_7b", device_map="auto", load_in_4bit=True
... )
```
### Generierte Ausgabe ist zu kurz/lang
Wenn in der Datei [~generation.GenerationConfig`] nichts angegeben ist, gibt `generate` standardmรครig bis zu 20 Token zurรผck. Wir empfehlen dringend, `max_new_tokens` in Ihrem `generate`-Aufruf manuell zu setzen, um die maximale Anzahl neuer Token zu kontrollieren, die zurรผckgegeben werden kรถnnen. Beachten Sie, dass LLMs (genauer gesagt, [decoder-only models](https://huggingface.co/learn/nlp-course/chapter1/6?fw=pt)) auch die Eingabeaufforderung als Teil der Ausgabe zurรผckgeben.
```py
>>> model_inputs = tokenizer(["A sequence of numbers: 1, 2"], return_tensors="pt").to("cuda")
>>> # By default, the output will contain up to 20 tokens
>>> generated_ids = model.generate(**model_inputs)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
'A sequence of numbers: 1, 2, 3, 4, 5'
>>> # Setting `max_new_tokens` allows you to control the maximum length
>>> generated_ids = model.generate(**model_inputs, max_new_tokens=50)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
'A sequence of numbers: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,'
```
### Falscher Generierungsmodus
Standardmรครig und sofern nicht in der Datei [~generation.GenerationConfig`] angegeben, wรคhlt `generate` bei jeder Iteration das wahrscheinlichste Token aus (gierige Dekodierung). Je nach Aufgabe kann dies unerwรผnscht sein; kreative Aufgaben wie Chatbots oder das Schreiben eines Aufsatzes profitieren vom Sampling. Andererseits profitieren Aufgaben, bei denen es auf die Eingabe ankommt, wie z.B. Audiotranskription oder รbersetzung, von der gierigen Dekodierung. Aktivieren Sie das Sampling mit `do_sample=True`. Mehr zu diesem Thema erfahren Sie in diesem [Blogbeitrag] (https://huggingface.co/blog/how-to-generate).
```py
>>> # Set seed or reproducibility -- you don't need this unless you want full reproducibility
>>> from transformers import set_seed
>>> set_seed(0)
>>> model_inputs = tokenizer(["I am a cat."], return_tensors="pt").to("cuda")
>>> # LLM + greedy decoding = repetitive, boring output
>>> generated_ids = model.generate(**model_inputs)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
'I am a cat. I am a cat. I am a cat. I am a cat'
>>> # With sampling, the output becomes more creative!
>>> generated_ids = model.generate(**model_inputs, do_sample=True)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
'I am a cat.\nI just need to be. I am always.\nEvery time'
```
### Falsche Auffรผllseite
LLMs sind [decoder-only](https://huggingface.co/learn/nlp-course/chapter1/6?fw=pt)-Architekturen, d.h. sie iterieren weiter รผber Ihre Eingabeaufforderung. Wenn Ihre Eingaben nicht die gleiche Lรคnge haben, mรผssen sie aufgefรผllt werden. Da LLMs nicht darauf trainiert sind, mit aufgefรผllten Token fortzufahren, muss Ihre Eingabe links aufgefรผllt werden. Vergessen Sie auch nicht, die Aufmerksamkeitsmaske an generate zu รผbergeben!
```py
>>> # The tokenizer initialized above has right-padding active by default: the 1st sequence,
>>> # which is shorter, has padding on the right side. Generation fails.
>>> model_inputs = tokenizer(
... ["1, 2, 3", "A, B, C, D, E"], padding=True, return_tensors="pt"
... ).to("cuda")
>>> generated_ids = model.generate(**model_inputs)
>>> tokenizer.batch_decode(generated_ids[0], skip_special_tokens=True)[0]
''
>>> # With left-padding, it works as expected!
>>> tokenizer = AutoTokenizer.from_pretrained("openlm-research/open_llama_7b", padding_side="left")
>>> tokenizer.pad_token = tokenizer.eos_token # Llama has no pad token by default
>>> model_inputs = tokenizer(
... ["1, 2, 3", "A, B, C, D, E"], padding=True, return_tensors="pt"
... ).to("cuda")
>>> generated_ids = model.generate(**model_inputs)
>>> tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
'1, 2, 3, 4, 5, 6,'
```
<!-- TODO: when the prompting guide is ready, mention the importance of setting the right prompt in this section -->
## Weitere Ressourcen
Wรคhrend der Prozess der autoregressiven Generierung relativ einfach ist, kann die optimale Nutzung Ihres LLM ein schwieriges Unterfangen sein, da es viele bewegliche Teile gibt. Fรผr Ihre nรคchsten Schritte, die Ihnen helfen, tiefer in die LLM-Nutzung und das Verstรคndnis einzutauchen:
<!-- TODO: mit neuen Anleitungen vervollstรคndigen -->
### Fortgeschrittene Nutzung generieren
1. [Leitfaden](generation_strategies) zur Steuerung verschiedener Generierungsmethoden, zur Einrichtung der Generierungskonfigurationsdatei und zum Streaming der Ausgabe;
2. API-Referenz zu [`~generation.GenerationConfig`], [`~generation.GenerationMixin.generate`] und [generate-bezogene Klassen](internal/generation_utils).
### LLM-Ranglisten
1. [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard), das sich auf die Qualitรคt der Open-Source-Modelle konzentriert;
2. [Open LLM-Perf Leaderboard](https://huggingface.co/spaces/optimum/llm-perf-leaderboard), das sich auf den LLM-Durchsatz konzentriert.
### Latenz und Durchsatz
1. [Leitfaden](main_classes/quantization) zur dynamischen Quantisierung, der Ihnen zeigt, wie Sie Ihren Speicherbedarf drastisch reduzieren kรถnnen.
### Verwandte Bibliotheken
1. [text-generation-inference](https://github.com/huggingface/text-generation-inference), ein produktionsreifer Server fรผr LLMs;
2. [`optimum`](https://github.com/huggingface/optimum), eine Erweiterung von ๐ค Transformers, die fรผr bestimmte Hardware-Gerรคte optimiert.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/quicktour.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Schnellstart
[[open-in-colab]]
Mit ๐ค Transformers kรถnnen Sie sofort loslegen! Verwenden Sie die [`pipeline`] fรผr schnelle Inferenz und laden Sie schnell ein vortrainiertes Modell und einen Tokenizer mit einer [AutoClass](./model_doc/auto), um Ihre Text-, Bild- oder Audioaufgabe zu lรถsen.
<Tip>
Alle in der Dokumentation vorgestellten Codebeispiele haben oben links einen Umschalter fรผr PyTorch und TensorFlow. Wenn
nicht, wird erwartet, dass der Code fรผr beide Backends ohne รnderungen funktioniert.
</Tip>
## Pipeline
[`pipeline`] ist der einfachste Weg, ein vortrainiertes Modell fรผr eine bestimmte Aufgabe zu verwenden.
<Youtube id="tiZFewofSLM"/>
Die [`pipeline`] unterstรผtzt viele gรคngige Aufgaben:
**Text**:
* Stimmungsanalyse: Klassifizierung der Polaritรคt eines gegebenen Textes.
* Textgenerierung (auf Englisch): Generierung von Text aus einer gegebenen Eingabe.
* Name-Entity-Recognition (NER): Kennzeichnung jedes Worts mit der Entitรคt, die es reprรคsentiert (Person, Datum, Ort usw.).
* Beantwortung von Fragen: Extrahieren der Antwort aus dem Kontext, wenn ein gewisser Kontext und eine Frage gegeben sind.
* Fill-mask: Ausfรผllen von Lรผcken in einem Text mit maskierten Wรถrtern.
* Zusammenfassung: Erstellung einer Zusammenfassung einer langen Text- oder Dokumentensequenz.
* รbersetzung: รbersetzen eines Textes in eine andere Sprache.
* Merkmalsextraktion: Erstellen einer Tensordarstellung des Textes.
**Bild**:
* Bildklassifizierung: Klassifizierung eines Bildes.
* Bildsegmentierung: Klassifizierung jedes Pixels in einem Bild.
* Objekterkennung: Erkennen von Objekten innerhalb eines Bildes.
**Audio**:
* Audioklassifizierung: Zuweisung eines Labels zu einem bestimmten Audiosegment.
* Automatische Spracherkennung (ASR): Transkription von Audiodaten in Text.
<Tip>
Fรผr mehr Details รผber die [`pipeline`] und assoziierte Aufgaben, schauen Sie in die Dokumentation [hier](./main_classes/pipelines).
</Tip>
### Verwendung der Pipeline
Im folgenden Beispiel werden Sie die [`pipeline`] fรผr die Stimmungsanalyse verwenden.
Installieren Sie die folgenden Abhรคngigkeiten, falls Sie dies nicht bereits getan haben:
<frameworkcontent>
<pt>
```bash
pip install torch
```
</pt>
<tf>
```bash
pip install tensorflow
```
</tf>
</frameworkcontent>
Importieren sie die [`pipeline`] und spezifizieren sie die Aufgabe, welche sie lรถsen mรถchten:
```py
>>> from transformers import pipeline
>>> classifier = pipeline("sentiment-analysis")
```
Die Pipeline lรคdt ein standardmรครiges [vortrainiertes Modell] (https://huggingface.co/distilbert-base-uncased-finetuned-sst-2-english) und einen Tokenizer fรผr die Stimmungs-Analyse herunter und speichert sie. Jetzt kรถnnen Sie den "Klassifikator" auf Ihren Zieltext anwenden:
```py
>>> classifier("We are very happy to show you the ๐ค Transformers library.")
[{'label': 'POSITIVE', 'score': 0.9998}]
```
For more than one sentence, pass a list of sentences to the [`pipeline`] which returns a list of dictionaries:
```py
>>> results = classifier(["We are very happy to show you the ๐ค Transformers library.", "We hope you don't hate it."])
>>> for result in results:
... print(f"label: {result['label']}, with score: {round(result['score'], 4)}")
label: POSITIVE, with score: 0.9998
label: NEGATIVE, with score: 0.5309
```
Die [`pipeline`] kann auch รผber einen ganzen Datensatz iterieren. Starten wir mit der Installation der [๐ค Datasets](https://huggingface.co/docs/datasets/) Bibliothek:
```bash
pip install datasets
```
Erstellen wir eine [`pipeline`] mit der Aufgabe die wir lรถsen und dem Modell welches wir nutzen mรถchten.
```py
>>> import torch
>>> from transformers import pipeline
>>> speech_recognizer = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-base-960h")
```
Als nรคchstes laden wir den Datensatz (siehe ๐ค Datasets [Quick Start](https://huggingface.co/docs/datasets/quickstart) fรผr mehr Details) welches wir nutzen mรถchten. Zum Beispiel laden wir den [MInDS-14](https://huggingface.co/datasets/PolyAI/minds14) Datensatz:
```py
>>> from datasets import load_dataset, Audio
>>> dataset = load_dataset("PolyAI/minds14", name="en-US", split="train") # doctest: +IGNORE_RESULT
```
Wir mรผssen sicherstellen, dass die Abtastrate des Datensatzes der Abtastrate entspricht, mit der `facebook/wav2vec2-base-960h` trainiert wurde.
```py
>>> dataset = dataset.cast_column("audio", Audio(sampling_rate=speech_recognizer.feature_extractor.sampling_rate))
```
Audiodateien werden automatisch geladen und neu abgetastet, wenn die Spalte "audio" aufgerufen wird.
Extrahieren wir die rohen Wellenform-Arrays der ersten 4 Beispiele und รผbergeben wir sie als Liste an die Pipeline:
```py
>>> result = speech_recognizer(dataset[:4]["audio"])
>>> print([d["text"] for d in result])
['I WOULD LIKE TO SET UP A JOINT ACCOUNT WITH MY PARTNER HOW DO I PROCEED WITH DOING THAT', "FODING HOW I'D SET UP A JOIN TO HET WITH MY WIFE AND WHERE THE AP MIGHT BE", "I I'D LIKE TOY SET UP A JOINT ACCOUNT WITH MY PARTNER I'M NOT SEEING THE OPTION TO DO IT ON THE AP SO I CALLED IN TO GET SOME HELP CAN I JUST DO IT OVER THE PHONE WITH YOU AND GIVE YOU THE INFORMATION OR SHOULD I DO IT IN THE AP AND I'M MISSING SOMETHING UQUETTE HAD PREFERRED TO JUST DO IT OVER THE PHONE OF POSSIBLE THINGS", 'HOW DO I THURN A JOIN A COUNT']
```
Bei einem grรถรeren Datensatz mit vielen Eingaben (wie bei Sprache oder Bildverarbeitung) sollten Sie einen Generator anstelle einer Liste รผbergeben, der alle Eingaben in den Speicher lรคdt. Weitere Informationen finden Sie in der [Pipeline-Dokumentation](./main_classes/pipelines).
### Ein anderes Modell und einen anderen Tokenizer in der Pipeline verwenden
Die [`pipeline`] kann jedes Modell aus dem [Model Hub] (https://huggingface.co/models) verwenden, wodurch es einfach ist, die [`pipeline`] fรผr andere Anwendungsfรคlle anzupassen. Wenn Sie beispielsweise ein Modell wรผnschen, das franzรถsischen Text verarbeiten kann, verwenden Sie die Tags im Model Hub, um nach einem geeigneten Modell zu filtern. Das oberste gefilterte Ergebnis liefert ein mehrsprachiges [BERT-Modell](https://huggingface.co/nlptown/bert-base-multilingual-uncased-sentiment), das auf die Stimmungsanalyse abgestimmt ist. Groรartig, verwenden wir dieses Modell!
```py
>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
```
<frameworkcontent>
<pt>
Use the [`AutoModelForSequenceClassification`] and [`AutoTokenizer`] to load the pretrained model and it's associated tokenizer (more on an `AutoClass` below):
```py
>>> from transformers import AutoTokenizer, AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained(model_name)
>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
```
</pt>
<tf>
Use the [`TFAutoModelForSequenceClassification`] and [`AutoTokenizer`] to load the pretrained model and it's associated tokenizer (more on an `TFAutoClass` below):
```py
>>> from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
>>> model = TFAutoModelForSequenceClassification.from_pretrained(model_name)
>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
```
</tf>
</frameworkcontent>
Dann kรถnnen Sie das Modell und den Tokenizer in der [`pipeline`] angeben und den `Klassifikator` auf Ihren Zieltext anwenden:
```py
>>> classifier = pipeline("sentiment-analysis", model=model, tokenizer=tokenizer)
>>> classifier("Nous sommes trรจs heureux de vous prรฉsenter la bibliothรจque ๐ค Transformers.")
[{'label': '5 stars', 'score': 0.7273}]
```
Wenn Sie kein Modell fรผr Ihren Anwendungsfall finden kรถnnen, mรผssen Sie ein vortrainiertes Modell auf Ihren Daten feinabstimmen. Schauen Sie sich unser [Feinabstimmungs-Tutorial](./training) an, um zu erfahren, wie das geht. Und schlieรlich, nachdem Sie Ihr trainiertes Modell verfeinert haben, sollten Sie es mit der Community im Model Hub teilen (siehe Tutorial [hier](./model_sharing)), um NLP fรผr alle zu demokratisieren! ๐ค
## AutoClass
<Youtube id="AhChOFRegn4"/>
Unter der Haube arbeiten die Klassen [`AutoModelForSequenceClassification`] und [`AutoTokenizer`] zusammen, um die [`pipeline`] zu betreiben. Eine [`AutoClass`](./model_doc/auto) ist eine Abkรผrzung, die automatisch die Architektur eines trainierten Modells aus dessen Namen oder Pfad abruft. Sie mรผssen nur die passende `AutoClass` fรผr Ihre Aufgabe und den zugehรถrigen Tokenizer mit [`AutoTokenizer`] auswรคhlen.
Kehren wir zu unserem Beispiel zurรผck und sehen wir uns an, wie Sie die `AutoClass` verwenden kรถnnen, um die Ergebnisse der [`pipeline`] zu replizieren.
### AutoTokenizer
Ein Tokenizer ist fรผr die Vorverarbeitung von Text in ein fรผr das Modell verstรคndliches Format zustรคndig. Zunรคchst zerlegt der Tokenisierer den Text in Wรถrter, die *Token* genannt werden. Es gibt mehrere Regeln fรผr den Tokenisierungsprozess, z. B. wie und auf welcher Ebene ein Wort aufgespalten wird (weitere Informationen รผber Tokenisierung [hier](./tokenizer_summary)). Das Wichtigste ist jedoch, dass Sie den Tokenizer mit demselben Modellnamen instanziieren mรผssen, um sicherzustellen, dass Sie dieselben Tokenisierungsregeln verwenden, mit denen ein Modell zuvor trainiert wurde.
Laden sie einen Tokenizer mit [`AutoTokenizer`]:
```py
>>> from transformers import AutoTokenizer
>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
```
Anschlieรend wandelt der Tokenizer die Token in Zahlen um, um einen Tensor als Eingabe fรผr das Modell zu konstruieren. Dieser wird als *Vokabular* des Modells bezeichnet.
รbergeben Sie Ihren Text an den Tokenizer:
```py
>>> encoding = tokenizer("We are very happy to show you the ๐ค Transformers library.")
>>> print(encoding)
{'input_ids': [101, 11312, 10320, 12495, 19308, 10114, 11391, 10855, 10103, 100, 58263, 13299, 119, 102],
'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}
```
Der Tokenizer gibt ein Wรถrterbuch zurรผck, das Folgendes enthรคlt:
* [input_ids](./glossary#input-ids): numerische Reprรคsentationen Ihrer Token.
* [atttention_mask](.glossary#attention-mask): gibt an, welche Token beachtet werden sollen.
Genau wie die [`pipeline`] akzeptiert der Tokenizer eine Liste von Eingaben. Darรผber hinaus kann der Tokenizer den Text auch auffรผllen und kรผrzen, um einen Stapel mit einheitlicher Lรคnge zurรผckzugeben:
<frameworkcontent>
<pt>
```py
>>> pt_batch = tokenizer(
... ["We are very happy to show you the ๐ค Transformers library.", "We hope you don't hate it."],
... padding=True,
... truncation=True,
... max_length=512,
... return_tensors="pt",
... )
```
</pt>
<tf>
```py
>>> tf_batch = tokenizer(
... ["We are very happy to show you the ๐ค Transformers library.", "We hope you don't hate it."],
... padding=True,
... truncation=True,
... max_length=512,
... return_tensors="tf",
... )
```
</tf>
</frameworkcontent>
Lesen Sie das Tutorial [preprocessing](./preprocessing) fรผr weitere Details zur Tokenisierung.
### AutoModel
<frameworkcontent>
<pt>
๐ค Transformers bietet eine einfache und einheitliche Mรถglichkeit, vortrainierte Instanzen zu laden. Das bedeutet, dass Sie ein [`AutoModel`] laden kรถnnen, wie Sie einen [`AutoTokenizer`] laden wรผrden. Der einzige Unterschied ist die Auswahl des richtigen [`AutoModel`] fรผr die Aufgabe. Da Sie eine Text- oder Sequenzklassifizierung vornehmen, laden Sie [`AutoModelForSequenceClassification`]:
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
>>> pt_model = AutoModelForSequenceClassification.from_pretrained(model_name)
```
<Tip>
In der [Aufgabenzusammenfassung](./task_summary) steht, welche [AutoModel]-Klasse fรผr welche Aufgabe zu verwenden ist.
</Tip>
Jetzt kรถnnen Sie Ihren vorverarbeiteten Stapel von Eingaben direkt an das Modell รผbergeben. Sie mรผssen nur das Wรถrterbuch entpacken, indem Sie `**` hinzufรผgen:
```py
>>> pt_outputs = pt_model(**pt_batch)
```
Das Modell gibt die endgรผltigen Aktivierungen in dem Attribut "logits" aus. Wenden Sie die Softmax-Funktion auf die "logits" an, um die Wahrscheinlichkeiten zu erhalten:
```py
>>> from torch import nn
>>> pt_predictions = nn.functional.softmax(pt_outputs.logits, dim=-1)
>>> print(pt_predictions)
tensor([[0.0021, 0.0018, 0.0115, 0.2121, 0.7725],
[0.2084, 0.1826, 0.1969, 0.1755, 0.2365]], grad_fn=<SoftmaxBackward0>)
```
</pt>
<tf>
๐ค Transformers bietet eine einfache und einheitliche Methode zum Laden von vortrainierten Instanzen. Das bedeutet, dass Sie ein [`TFAutoModel`] genauso laden kรถnnen, wie Sie einen [`AutoTokenizer`] laden wรผrden. Der einzige Unterschied ist die Auswahl des richtigen [`TFAutoModel`] fรผr die Aufgabe. Da Sie Text - oder Sequenz - Klassifizierung machen, laden Sie [`TFAutoModelForSequenceClassification`]:
```py
>>> from transformers import TFAutoModelForSequenceClassification
>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained(model_name)
```
<Tip>
In der [Aufgabenzusammenfassung](./task_summary) steht, welche [AutoModel]-Klasse fรผr welche Aufgabe zu verwenden ist.
</Tip>
Jetzt kรถnnen Sie Ihren vorverarbeiteten Stapel von Eingaben direkt an das Modell รผbergeben, indem Sie die Wรถrterbuchschlรผssel direkt an die Tensoren รผbergeben:
```py
>>> tf_outputs = tf_model(tf_batch)
```
Das Modell gibt die endgรผltigen Aktivierungen in dem Attribut "logits" aus. Wenden Sie die Softmax-Funktion auf die "logits" an, um die Wahrscheinlichkeiten zu erhalten:
```py
>>> import tensorflow as tf
>>> tf_predictions = tf.nn.softmax(tf_outputs.logits, axis=-1)
>>> tf_predictions # doctest: +IGNORE_RESULT
```
</tf>
</frameworkcontent>
<Tip>
Alle ๐ค Transformers-Modelle (PyTorch oder TensorFlow) geben die Tensoren *vor* der endgรผltigen Aktivierungsfunktion
Funktion (wie Softmax) aus, da die endgรผltige Aktivierungsfunktion oft mit dem Verlusten verschmolzen ist.
</Tip>
Modelle sind ein standardmรครiges [`torch.nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) oder ein [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model), sodass Sie sie in Ihrer รผblichen Trainingsschleife verwenden kรถnnen. Um jedoch die Dinge einfacher zu machen, bietet ๐ค Transformers eine [`Trainer`]-Klasse fรผr PyTorch, die Funktionalitรคt fรผr verteiltes Training, gemischte Prรคzision und mehr bietet. Fรผr TensorFlow kรถnnen Sie die Methode `fit` aus [Keras](https://keras.io/) verwenden. Siehe das [training tutorial](./training) fรผr weitere Details.
<Tip>
Transformers-Modellausgaben sind spezielle Datenklassen, so dass ihre Attribute in einer IDE automatisch vervollstรคndigt werden.
Die Modellausgรคnge verhalten sich auch wie ein Tupel oder ein Wรถrterbuch (z.B. kรถnnen Sie mit einem Integer, einem Slice oder einem String indexieren), wobei die Attribute, die "None" sind, ignoriert werden.
</Tip>
### Modell speichern
<frameworkcontent>
<pt>
Sobald Ihr Modell feinabgestimmt ist, kรถnnen Sie es mit seinem Tokenizer speichern, indem Sie [`PreTrainedModel.save_pretrained`] verwenden:
```py
>>> pt_save_directory = "./pt_save_pretrained"
>>> tokenizer.save_pretrained(pt_save_directory) # doctest: +IGNORE_RESULT
>>> pt_model.save_pretrained(pt_save_directory)
```
Wenn Sie bereit sind, das Modell erneut zu verwenden, laden Sie es mit [`PreTrainedModel.from_pretrained`]:
```py
>>> pt_model = AutoModelForSequenceClassification.from_pretrained("./pt_save_pretrained")
```
</pt>
<tf>
Sobald Ihr Modell feinabgestimmt ist, kรถnnen Sie es mit seinem Tokenizer unter Verwendung von [`TFPreTrainedModel.save_pretrained`] speichern:
```py
>>> tf_save_directory = "./tf_save_pretrained"
>>> tokenizer.save_pretrained(tf_save_directory) # doctest: +IGNORE_RESULT
>>> tf_model.save_pretrained(tf_save_directory)
```
Wenn Sie bereit sind, das Modell wieder zu verwenden, laden Sie es mit [`TFPreTrainedModel.from_pretrained`]:
```py
>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained("./tf_save_pretrained")
```
</tf>
</frameworkcontent>
Ein besonders cooles ๐ค Transformers-Feature ist die Mรถglichkeit, ein Modell zu speichern und es entweder als PyTorch- oder TensorFlow-Modell wieder zu laden. Der Parameter "from_pt" oder "from_tf" kann das Modell von einem Framework in das andere konvertieren:
<frameworkcontent>
<pt>
```py
>>> from transformers import AutoModel
>>> tokenizer = AutoTokenizer.from_pretrained(tf_save_directory)
>>> pt_model = AutoModelForSequenceClassification.from_pretrained(tf_save_directory, from_tf=True)
```
</pt>
<tf>
```py
>>> from transformers import TFAutoModel
>>> tokenizer = AutoTokenizer.from_pretrained(pt_save_directory)
>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained(pt_save_directory, from_pt=True)
```
</tf>
</frameworkcontent>
## Custom model builds
Sie kรถnnen die Konfigurationsklasse des Modells รคndern, um zu bestimmen, wie ein Modell aufgebaut ist. Die Konfiguration legt die Attribute eines Modells fest, z. B. die Anzahl der verborgenen Schichten oder der Aufmerksamkeitskรถpfe. Wenn Sie ein Modell aus einer benutzerdefinierten Konfigurationsklasse initialisieren, beginnen Sie bei Null. Die Modellattribute werden zufรคllig initialisiert, und Sie mรผssen das Modell trainieren, bevor Sie es verwenden kรถnnen, um aussagekrรคftige Ergebnisse zu erhalten.
Beginnen Sie mit dem Import von [`AutoConfig`] und laden Sie dann das trainierte Modell, das Sie รคndern mรถchten. Innerhalb von [`AutoConfig.from_pretrained`] kรถnnen Sie das Attribut angeben, das Sie รคndern mรถchten, z. B. die Anzahl der Aufmerksamkeitskรถpfe:
```py
>>> from transformers import AutoConfig
>>> my_config = AutoConfig.from_pretrained("distilbert-base-uncased", n_heads=12)
```
<frameworkcontent>
<pt>
Create a model from your custom configuration with [`AutoModel.from_config`]:
```py
>>> from transformers import AutoModel
>>> my_model = AutoModel.from_config(my_config)
```
</pt>
<tf>
Create a model from your custom configuration with [`TFAutoModel.from_config`]:
```py
>>> from transformers import TFAutoModel
>>> my_model = TFAutoModel.from_config(my_config)
```
</tf>
</frameworkcontent>
Weitere Informationen zur Erstellung von benutzerdefinierten Konfigurationen finden Sie in der Anleitung [Erstellen einer benutzerdefinierten Architektur](./create_a_model).
## Wie geht es weiter?
Nachdem Sie nun die ๐ค Transformers-Kurztour abgeschlossen haben, schauen Sie sich unsere Anleitungen an und erfahren Sie, wie Sie spezifischere Dinge tun kรถnnen, wie das Schreiben eines benutzerdefinierten Modells, die Feinabstimmung eines Modells fรผr eine Aufgabe und wie man ein Modell mit einem Skript trainiert. Wenn Sie mehr รผber die Kernkonzepte von ๐ค Transformers erfahren mรถchten, nehmen Sie sich eine Tasse Kaffee und werfen Sie einen Blick auf unsere konzeptionellen Leitfรคden!
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/de/testing.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Testen
Werfen wir einen Blick darauf, wie ๐ค Transformers-Modelle getestet werden und wie Sie neue Tests schreiben und die vorhandenen verbessern kรถnnen.
Es gibt 2 Testsuiten im Repository:
1. `tests` -- Tests fรผr die allgemeine API
2. `examples` -- Tests hauptsรคchlich fรผr verschiedene Anwendungen, die nicht Teil der API sind
## Wie Transformatoren getestet werden
1. Sobald ein PR eingereicht wurde, wird er mit 9 CircleCi Jobs getestet. Jeder neue Commit zu diesem PR wird erneut getestet. Diese Auftrรคge
sind in dieser [Konfigurationsdatei](https://github.com/huggingface/transformers/tree/main/.circleci/config.yml) definiert, so dass Sie bei Bedarf die gleiche Umgebung auf Ihrem Rechner reproduzieren kรถnnen.
Umgebung auf Ihrem Rechner reproduzieren kรถnnen.
Diese CI-Jobs fรผhren keine `@slow`-Tests durch.
2. Es gibt 3 Jobs, die von [github actions](https://github.com/huggingface/transformers/actions) ausgefรผhrt werden:
- [torch hub integration](https://github.com/huggingface/transformers/tree/main/.github/workflows/github-torch-hub.yml): prรผft, ob die torch hub
Integration funktioniert.
- [self-hosted (push)](https://github.com/huggingface/transformers/tree/main/.github/workflows/self-push.yml): fรผhrt schnelle Tests auf der GPU nur bei Commits auf
`main`. Es wird nur ausgefรผhrt, wenn ein Commit auf `main` den Code in einem der folgenden Ordner aktualisiert hat: `src`,
`tests`, `.github` (um zu verhindern, dass er auf hinzugefรผgten Modellkarten, Notebooks usw. lรคuft)
- [self-hosted runner](https://github.com/huggingface/transformers/tree/main/.github/workflows/self-scheduled.yml): fรผhrt normale und langsame Tests auf GPU in
`tests` und `examples`:
```bash
RUN_SLOW=1 pytest tests/
RUN_SLOW=1 pytest examples/
```
Die Ergebnisse kรถnnen Sie [hier](https://github.com/huggingface/transformers/actions) sehen.
## Tests ausfรผhren
### Auswahl der auszufรผhrenden Tests
In diesem Dokument wird ausfรผhrlich erlรคutert, wie Tests ausgefรผhrt werden kรถnnen. Wenn Sie nach der Lektรผre noch mehr Details benรถtigen
finden Sie diese [hier](https://docs.pytest.org/en/latest/usage.html).
Hier sind einige der nรผtzlichsten Mรถglichkeiten, Tests auszufรผhren.
Alle ausfรผhren:
```console
pytest
```
oder:
```bash
make test
```
Beachten Sie, dass Letzteres wie folgt definiert ist:
```bash
python -m pytest -n auto --dist=loadfile -s -v ./tests/
```
was pytest anweist:
- so viele Testprozesse laufen zu lassen, wie es CPU-Kerne gibt (was zu viele sein kรถnnten, wenn Sie nicht รผber eine Menge RAM verfรผgen!)
- sicherzustellen, dass alle Tests aus derselben Datei von demselben Testprozess ausgefรผhrt werden
- Erfassen Sie keine Ausgaben
- im ausfรผhrlichen Modus laufen lassen
### Abrufen der Liste aller Tests
Alle Tests der Testsuite:
```bash
pytest --collect-only -q
```
Alle Tests einer bestimmten Testdatei:
```bash
pytest tests/test_optimization.py --collect-only -q
```
### Fรผhren Sie ein bestimmtes Testmodul aus
Um ein einzelnes Testmodul auszufรผhren:
```bash
pytest tests/utils/test_logging.py
```
### Spezifische Tests ausfรผhren
Da unittest in den meisten Tests verwendet wird, mรผssen Sie, um bestimmte Untertests auszufรผhren, den Namen der unittest
Klasse, die diese Tests enthรคlt. Er kรถnnte zum Beispiel lauten:
```bash
pytest tests/test_optimization.py::OptimizationTest::test_adam_w
```
Hier:
- `tests/test_optimization.py` - die Datei mit den Tests
- `OptimizationTest` - der Name der Klasse
- `test_adam_w` - der Name der spezifischen Testfunktion
Wenn die Datei mehrere Klassen enthรคlt, kรถnnen Sie auswรคhlen, dass nur die Tests einer bestimmten Klasse ausgefรผhrt werden sollen. Zum Beispiel:
```bash
pytest tests/test_optimization.py::OptimizationTest
```
fรผhrt alle Tests innerhalb dieser Klasse aus.
Wie bereits erwรคhnt, kรถnnen Sie sehen, welche Tests in der Klasse `OptimizationTest` enthalten sind, indem Sie sie ausfรผhren:
```bash
pytest tests/test_optimization.py::OptimizationTest --collect-only -q
```
Sie kรถnnen Tests mit Hilfe von Schlรผsselwortausdrรผcken ausfรผhren.
Um nur Tests auszufรผhren, deren Name `adam` enthรคlt:
```bash
pytest -k adam tests/test_optimization.py
```
Die logischen `und` und `oder` kรถnnen verwendet werden, um anzugeben, ob alle Schlรผsselwรถrter รผbereinstimmen sollen oder nur eines. `nicht` kann verwendet werden, um
negieren.
Um alle Tests auszufรผhren, auรer denen, deren Name `adam` enthรคlt:
```bash
pytest -k "not adam" tests/test_optimization.py
```
Und Sie kรถnnen die beiden Muster in einem kombinieren:
```bash
pytest -k "ada and not adam" tests/test_optimization.py
```
Um zum Beispiel sowohl `test_adafactor` als auch `test_adam_w` auszufรผhren, kรถnnen Sie verwenden:
```bash
pytest -k "test_adam_w or test_adam_w" tests/test_optimization.py
```
Beachten Sie, dass wir hier `oder` verwenden, da wir wollen, dass eines der Schlรผsselwรถrter รผbereinstimmt, um beide einzuschlieรen.
Wenn Sie nur Tests einschlieรen mรถchten, die beide Muster enthalten, mรผssen Sie `und` verwenden:
```bash
pytest -k "test and ada" tests/test_optimization.py
```
### Fรผhren Sie `accelerate` Tests durch
Manchmal mรผssen Sie `accelerate` Tests fรผr Ihre Modelle ausfรผhren. Dazu fรผgen Sie einfach `-m accelerate_tests` zu Ihrem Befehl hinzu, wenn Sie diese Tests bei einem `OPT`-Lauf ausfรผhren mรถchten:
```bash
RUN_SLOW=1 pytest -m accelerate_tests tests/models/opt/test_modeling_opt.py
```
### Dokumentationstests ausfรผhren
Um zu testen, ob die Dokumentationsbeispiele korrekt sind, sollten Sie รผberprรผfen, ob die `doctests` erfolgreich sind.
Lassen Sie uns als Beispiel den docstring von [WhisperModel.forward](https://github.com/huggingface/transformers/blob/main/src/transformers/models/whisper/modeling_whisper.py#L1017-L1035) verwenden:
```python
r"""
Returns:
Example:
```python
>>> import torch
>>> from transformers import WhisperModel, WhisperFeatureExtractor
>>> from datasets import load_dataset
>>> model = WhisperModel.from_pretrained("openai/whisper-base")
>>> feature_extractor = WhisperFeatureExtractor.from_pretrained("openai/whisper-base")
>>> ds = load_dataset("hf-internal-testing/librispeech_asr_dummy", "clean", split="validation")
>>> inputs = feature_extractor(ds[0]["audio"]["array"], return_tensors="pt")
>>> input_features = inputs.input_features
>>> decoder_input_ids = torch.tensor([[1, 1]]) * model.config.decoder_start_token_id
>>> last_hidden_state = model(input_features, decoder_input_ids=decoder_input_ids).last_hidden_state
>>> list(last_hidden_state.shape)
[1, 2, 512]
```"""
```
Fรผhren Sie einfach die folgende Zeile aus, um automatisch jedes docstring-Beispiel in der gewรผnschten Datei zu testen:
```bash
pytest --doctest-modules <path_to_file_or_dir>
```
Wenn die Datei eine Markdown-Erweiterung hat, sollten Sie das Argument `--doctest-glob="*.md"` hinzufรผgen.
### Nur geรคnderte Tests ausfรผhren
Mit [pytest-picked](https://github.com/anapaulagomes/pytest-picked) kรถnnen Sie die Tests ausfรผhren, die sich auf die unstaged Dateien oder den aktuellen Zweig (gemรคร Git) beziehen. Auf diese Weise kรถnnen Sie schnell testen, ob Ihre รnderungen nichts kaputt gemacht haben.
nichts kaputt gemacht haben, da die Tests fรผr Dateien, die Sie nicht verรคndert haben, nicht ausgefรผhrt werden.
```bash
pip install pytest-picked
```
```bash
pytest --picked
```
Alle Tests werden von Dateien und Ordnern ausgefรผhrt, die geรคndert, aber noch nicht รผbergeben wurden.
### Fehlgeschlagene Tests bei รnderung der Quelle automatisch wiederholen
[pytest-xdist](https://github.com/pytest-dev/pytest-xdist) bietet eine sehr nรผtzliche Funktion zur Erkennung aller fehlgeschlagenen
Tests zu erkennen und dann darauf zu warten, dass Sie Dateien รคndern, um die fehlgeschlagenen Tests so lange zu wiederholen, bis sie erfolgreich sind, wรคhrend Sie die
sie reparieren. So mรผssen Sie pytest nicht erneut starten, nachdem Sie die Korrektur vorgenommen haben. Dies wird so lange wiederholt, bis alle Tests bestanden sind.
Danach wird erneut ein vollstรคndiger Durchlauf durchgefรผhrt.
```bash
pip install pytest-xdist
```
So rufen Sie den Modus auf: `pytest -f` oder `pytest --looponfail`
Datei-รnderungen werden erkannt, indem die Wurzelverzeichnisse von `looponfailroots` und alle ihre Inhalte (rekursiv) untersucht werden.
Wenn die Vorgabe fรผr diesen Wert fรผr Sie nicht funktioniert, kรถnnen Sie ihn in Ihrem Projekt รคndern, indem Sie eine Konfigurations
Option in der Datei `setup.cfg` รคndern:
```ini
[tool:pytest]
looponfailroots = transformers tests
```
oder die Dateien `pytest.ini`/`tox.ini``:
```ini
[pytest]
looponfailroots = transformers tests
```
Dies wรผrde dazu fรผhren, dass nur nach Dateiรคnderungen in den jeweiligen Verzeichnissen gesucht wird, die relativ zum Verzeichnis der ini-Datei angegeben sind.
Verzeichnis.
[pytest-watch](https://github.com/joeyespo/pytest-watch) ist eine alternative Implementierung dieser Funktionalitรคt.
### รberspringen eines Testmoduls
Wenn Sie alle Testmodule ausfรผhren mรถchten, mit Ausnahme einiger weniger, kรถnnen Sie diese ausschlieรen, indem Sie eine explizite Liste der auszufรผhrenden Tests angeben. Fรผr
Beispiel: Um alle Tests auรer `test_modeling_*.py` auszufรผhren:
```bash
pytest *ls -1 tests/*py | grep -v test_modeling*
```
### Status leeren
CI-Builds und wenn Isolation wichtig ist (gegen Geschwindigkeit), sollte der Cache geleert werden:
```bash
pytest --cache-clear tests
```
### Tests parallel ausfรผhren
Wie bereits erwรคhnt, fรผhrt `make test` รผber das Plugin `pytest-xdist` Tests parallel aus (Argument `-n X`, z.B. `-n 2`
um 2 Jobs parallel laufen zu lassen).
Mit der Option `--dist=` von `pytest-xdist` kรถnnen Sie steuern, wie die Tests gruppiert werden. Mit `--dist=loadfile` werden die
Tests, die sich in einer Datei befinden, in denselben Prozess.
Da die Reihenfolge der ausgefรผhrten Tests unterschiedlich und nicht vorhersehbar ist, kann die Ausfรผhrung der Testsuite mit `pytest-xdist`
zu Fehlern fรผhrt (was bedeutet, dass wir einige unentdeckte gekoppelte Tests haben), verwenden Sie [pytest-replay](https://github.com/ESSS/pytest-replay), um die Tests in der gleichen Reihenfolge abzuspielen, was dabei helfen sollte
diese fehlgeschlagene Sequenz auf ein Minimum zu reduzieren.
### Testreihenfolge und Wiederholung
Es ist gut, die Tests mehrmals zu wiederholen, nacheinander, zufรคllig oder in Gruppen, um mรถgliche
Abhรคngigkeiten und zustandsbezogene Fehler zu erkennen (Abriss). Und die einfache, mehrfache Wiederholung ist einfach gut, um
einige Probleme zu erkennen, die durch die Zufรคlligkeit von DL aufgedeckt werden.
#### Wiederholungstests
- [pytest-flakefinder](https://github.com/dropbox/pytest-flakefinder):
```bash
pip install pytest-flakefinder
```
Und fรผhren Sie dann jeden Test mehrmals durch (standardmรครig 50):
```bash
pytest --flake-finder --flake-runs=5 tests/test_failing_test.py
```
<Tip>
Dieses Plugin funktioniert nicht mit dem `-n` Flag von `pytest-xdist`.
</Tip>
<Tip>
Es gibt noch ein anderes Plugin `pytest-repeat`, aber es funktioniert nicht mit `unittest`.
</Tip>
#### Run tests in a random order
```bash
pip install pytest-random-order
```
Wichtig: Das Vorhandensein von `pytest-random-order` sorgt fรผr eine automatische Zufallsanordnung der Tests, es sind keine Konfigurationsรคnderungen oder
Befehlszeilenoptionen sind nicht erforderlich.
Wie bereits erlรคutert, ermรถglicht dies die Erkennung von gekoppelten Tests - bei denen der Zustand eines Tests den Zustand eines anderen beeinflusst. Wenn
`pytest-random-order` installiert ist, gibt es den Zufallswert aus, der fรผr diese Sitzung verwendet wurde, z.B:
```bash
pytest tests
[...]
Using --random-order-bucket=module
Using --random-order-seed=573663
```
Wenn eine bestimmte Sequenz fehlschlรคgt, kรถnnen Sie sie reproduzieren, indem Sie genau diesen Seed hinzufรผgen, z.B:
```bash
pytest --random-order-seed=573663
[...]
Using --random-order-bucket=module
Using --random-order-seed=573663
```
Es wird nur dann die exakte Reihenfolge reproduzieren, wenn Sie genau dieselbe Liste von Tests (oder gar keine Liste) verwenden. Sobald Sie beginnen, die Liste
die Liste manuell einzugrenzen, kรถnnen Sie sich nicht mehr auf den Seed verlassen, sondern mรผssen die Tests manuell in der genauen Reihenfolge auflisten
auflisten und pytest anweisen, sie nicht zu randomisieren, indem Sie `--random-order-bucket=none` verwenden, z.B.:
```bash
pytest --random-order-bucket=none tests/test_a.py tests/test_c.py tests/test_b.py
```
So deaktivieren Sie das Shuffling fรผr alle Tests:
```bash
pytest --random-order-bucket=none
```
Standardmรครig ist `--random-order-bucket=module` impliziert, wodurch die Dateien auf den Modulebenen gemischt werden. Es kann auch
auf den Ebenen `class`, `package`, `global` und `none` mischen. Die vollstรคndigen Details entnehmen Sie bitte der
[Dokumentation] (https://github.com/jbasko/pytest-random-order).
Eine weitere Alternative zur Randomisierung ist: [`pytest-random`](https://github.com/pytest-dev/pytest-randomly). Dieses
Modul hat eine sehr รคhnliche Funktionalitรคt/Schnittstelle, aber es hat nicht die Eimermodi, die in
`pytest-random-order` zur Verfรผgung. Es hat das gleiche Problem, dass es sich nach der Installation aufdrรคngt.
### Variationen von Aussehen und Bedienung
#### pytest-zucker
[pytest-sugar](https://github.com/Frozenball/pytest-sugar) ist ein Plugin, das das Erscheinungsbild verbessert, eine
Fortschrittsbalken hinzufรผgt und Tests, die fehlschlagen, sowie die Bestรคtigung sofort anzeigt. Es wird bei der Installation automatisch aktiviert.
```bash
pip install pytest-sugar
```
Um Tests ohne sie durchzufรผhren, fรผhren Sie aus:
```bash
pytest -p no:sugar
```
oder deinstallieren Sie es.
#### Melden Sie den Namen jedes Subtests und seinen Fortschritt
Fรผr einen einzelnen oder eine Gruppe von Tests รผber `pytest` (nach `pip install pytest-pspec`):
```bash
pytest --pspec tests/test_optimization.py
```
#### Zeigt fehlgeschlagene Tests sofort an
[pytest-instafail](https://github.com/pytest-dev/pytest-instafail) zeigt Fehlschlรคge und Fehler sofort an, anstatt
bis zum Ende der Testsitzung zu warten.
```bash
pip install pytest-instafail
```
```bash
pytest --instafail
```
### Zu GPU oder nicht zu GPU
Bei einem GPU-aktivierten Setup fรผgen Sie zum Testen im reinen CPU-Modus `CUDA_VISIBLE_DEVICES=""` hinzu:
```bash
CUDA_VISIBLE_DEVICES="" pytest tests/utils/test_logging.py
```
oder wenn Sie mehrere Grafikprozessoren haben, kรถnnen Sie angeben, welcher von `pytest` verwendet werden soll. Wenn Sie zum Beispiel nur den
zweiten Grafikkarte zu verwenden, wenn Sie die Grafikkarten `0` und `1` haben, kรถnnen Sie folgendes ausfรผhren:
```bash
CUDA_VISIBLE_DEVICES="1" pytest tests/utils/test_logging.py
```
Dies ist praktisch, wenn Sie verschiedene Aufgaben auf verschiedenen GPUs ausfรผhren mรถchten.
Einige Tests mรผssen nur auf der CPU ausgefรผhrt werden, andere entweder auf der CPU, der GPU oder der TPU und wieder andere auf mehreren GPUs. Die folgenden skip
Dekorateure werden verwendet, um die Anforderungen von Tests in Bezug auf CPU/GPU/TPU festzulegen:
- `require_torch` - dieser Test wird nur unter Torch ausgefรผhrt
- `require_torch_gpu` - wie `require_torch` plus erfordert mindestens 1 GPU
- `require_torch_multi_gpu` - wie `require_torch` und zusรคtzlich mindestens 2 GPUs erforderlich
- `require_torch_non_multi_gpu` - wie `require_torch` plus benรถtigt 0 oder 1 GPUs
- `require_torch_up_to_2_gpus` - wie `require_torch` plus erfordert 0 oder 1 oder 2 GPUs
- `require_torch_tpu` - wie `require_torch` plus erfordert mindestens 1 TPU
Lassen Sie uns die GPU-Anforderungen in der folgenden Tabelle darstellen:
| n gpus | decorator |
|--------|--------------------------------|
| `>= 0` | `@require_torch` |
| `>= 1` | `@require_torch_gpu` |
| `>= 2` | `@require_torch_multi_gpu` |
| `< 2` | `@require_torch_non_multi_gpu` |
| `< 3` | `@require_torch_up_to_2_gpus` |
Hier ist zum Beispiel ein Test, der nur ausgefรผhrt werden muss, wenn 2 oder mehr GPUs verfรผgbar sind und pytorch installiert ist:
```python no-style
@require_torch_multi_gpu
def test_example_with_multi_gpu():
```
Wenn ein Test `tensorflow` benรถtigt, verwenden Sie den Dekorator `require_tf`. Zum Beispiel:
```python no-style
@require_tf
def test_tf_thing_with_tensorflow():
```
Diese Dekors kรถnnen gestapelt werden. Wenn zum Beispiel ein Test langsam ist und mindestens eine GPU unter pytorch benรถtigt, kรถnnen Sie
wie Sie ihn einrichten kรถnnen:
```python no-style
@require_torch_gpu
@slow
def test_example_slow_on_gpu():
```
Einige Dekoratoren wie `@parametrized` schreiben Testnamen um, daher mรผssen `@require_*`-Sprungdekoratoren als letztes aufgefรผhrt werden.
zuletzt aufgefรผhrt werden, damit sie korrekt funktionieren. Hier ist ein Beispiel fรผr die korrekte Verwendung:
```python no-style
@parameterized.expand(...)
@require_torch_multi_gpu
def test_integration_foo():
```
Dieses Problem mit der Reihenfolge gibt es bei `@pytest.mark.parametrize` nicht, Sie kรถnnen es an den Anfang oder an den Schluss setzen und es wird trotzdem funktionieren.
funktionieren. Aber es funktioniert nur bei Nicht-Unittests.
Innerhalb von Tests:
- Wie viele GPUs sind verfรผgbar:
```python
from transformers.testing_utils import get_gpu_count
n_gpu = get_gpu_count() # works with torch and tf
```
### Testen mit einem bestimmten PyTorch-Backend oder Gerรคt
Um die Testsuite auf einem bestimmten Torch-Gerรคt auszufรผhren, fรผgen Sie `TRANSFORMERS_TEST_DEVICE="$Gerรคt"` hinzu, wobei `$Gerรคt` das Ziel-Backend ist. Zum Beispiel, um nur auf der CPU zu testen:
```bash
TRANSFORMERS_TEST_DEVICE="cpu" pytest tests/utils/test_logging.py
```
Diese Variable ist nรผtzlich, um benutzerdefinierte oder weniger verbreitete PyTorch-Backends wie `mps` zu testen. Sie kann auch verwendet werden, um den gleichen Effekt wie `CUDA_VISIBLE_DEVICES` zu erzielen, indem Sie bestimmte GPUs anvisieren oder im reinen CPU-Modus testen.
Bestimmte Gerรคte erfordern einen zusรคtzlichen Import, nachdem Sie `torch` zum ersten Mal importiert haben. Dies kann รผber die Umgebungsvariable `TRANSFORMERS_TEST_BACKEND` festgelegt werden:
```bash
TRANSFORMERS_TEST_BACKEND="torch_npu" pytest tests/utils/test_logging.py
```
### Verteiltes Training
`pytest` kann nicht direkt mit verteiltem Training umgehen. Wenn dies versucht wird, tun die Unterprozesse nicht das Richtige
und denken am Ende, sie seien `pytest` und beginnen, die Testsuite in Schleifen auszufรผhren. Es funktioniert jedoch, wenn man
einen normalen Prozess erzeugt, der dann mehrere Worker erzeugt und die IO-Pipes verwaltet.
Hier sind einige Tests, die dies verwenden:
- [test_trainer_distributed.py](https://github.com/huggingface/transformers/tree/main/tests/trainer/test_trainer_distributed.py)
- [test_deepspeed.py](https://github.com/huggingface/transformers/tree/main/tests/deepspeed/test_deepspeed.py)
Um direkt mit der Ausfรผhrung zu beginnen, suchen Sie in diesen Tests nach dem Aufruf `execute_subprocess_async`.
Sie benรถtigen mindestens 2 GPUs, um diese Tests in Aktion zu sehen:
```bash
CUDA_VISIBLE_DEVICES=0,1 RUN_SLOW=1 pytest -sv tests/test_trainer_distributed.py
```
### Erfassung von Ausgaben
Wรคhrend der Testausfรผhrung werden alle Ausgaben, die an `stdout` und `stderr` gesendet werden, aufgezeichnet. Wenn ein Test oder eine Setup-Methode fehlschlรคgt, wird die
wird die entsprechende aufgezeichnete Ausgabe in der Regel zusammen mit dem Fehler-Traceback angezeigt.
Um die Aufzeichnung von Ausgaben zu deaktivieren und `stdout` und `stderr` normal zu erhalten, verwenden Sie `-s` oder `--capture=no`:
```bash
pytest -s tests/utils/test_logging.py
```
So senden Sie Testergebnisse an die JUnit-Formatausgabe:
```bash
py.test tests --junitxml=result.xml
```
### Farbsteuerung
Keine Farbe zu haben (z.B. gelb auf weiรem Hintergrund ist nicht lesbar):
```bash
pytest --color=no tests/utils/test_logging.py
```
### Testbericht an den Online-Dienst pastebin senden
Erstellen Sie eine URL fรผr jeden Testfehler:
```bash
pytest --pastebin=failed tests/utils/test_logging.py
```
Dadurch werden Informationen รผber den Testlauf an einen entfernten Paste-Dienst รผbermittelt und eine URL fรผr jeden Fehlschlag bereitgestellt. Sie kรถnnen die
Tests wie gewohnt auswรคhlen oder z.B. -x hinzufรผgen, wenn Sie nur einen bestimmten Fehler senden mรถchten.
Erstellen einer URL fรผr ein ganzes Testsitzungsprotokoll:
```bash
pytest --pastebin=all tests/utils/test_logging.py
```
## Tests schreiben
๐ค Die Tests von Transformers basieren auf `unittest`, werden aber von `pytest` ausgefรผhrt, so dass die meiste Zeit Funktionen aus beiden Systemen
verwendet werden kรถnnen.
Sie kรถnnen [hier](https://docs.pytest.org/en/stable/unittest.html) nachlesen, welche Funktionen unterstรผtzt werden, aber das Wichtigste ist
Wichtig ist, dass die meisten `pytest`-Fixtures nicht funktionieren. Auch die Parametrisierung nicht, aber wir verwenden das Modul
`parametrisiert`, das auf รคhnliche Weise funktioniert.
### Parametrisierung
Oft besteht die Notwendigkeit, denselben Test mehrmals auszufรผhren, aber mit unterschiedlichen Argumenten. Das kรถnnte innerhalb des Tests geschehen
des Tests gemacht werden, aber dann gibt es keine Mรถglichkeit, den Test mit nur einem Satz von Argumenten auszufรผhren.
```python
# test_this1.py
import unittest
from parameterized import parameterized
class TestMathUnitTest(unittest.TestCase):
@parameterized.expand(
[
("negative", -1.5, -2.0),
("integer", 1, 1.0),
("large fraction", 1.6, 1),
]
)
def test_floor(self, name, input, expected):
assert_equal(math.floor(input), expected)
```
Nun wird dieser Test standardmรครig 3 Mal ausgefรผhrt, wobei jedes Mal die letzten 3 Argumente von `test_floor` den entsprechenden Argumenten in der Parameterliste zugeordnet werden.
die entsprechenden Argumente in der Parameterliste.
Sie kรถnnen auch nur die Parameter `negativ` und `ganzzahlig` mit ausfรผhren:
```bash
pytest -k "negative and integer" tests/test_mytest.py
```
oder alle Untertests auรer `negativ`, mit:
```bash
pytest -k "not negative" tests/test_mytest.py
```
Neben der Verwendung des gerade erwรคhnten Filters `-k` kรถnnen Sie auch den genauen Namen jedes Untertests herausfinden und jeden
oder alle unter Verwendung ihrer genauen Namen ausfรผhren.
```bash
pytest test_this1.py --collect-only -q
```
und es wird aufgelistet:
```bash
test_this1.py::TestMathUnitTest::test_floor_0_negative
test_this1.py::TestMathUnitTest::test_floor_1_integer
test_this1.py::TestMathUnitTest::test_floor_2_large_fraction
```
Jetzt kรถnnen Sie also nur 2 spezifische Untertests durchfรผhren:
```bash
pytest test_this1.py::TestMathUnitTest::test_floor_0_negative test_this1.py::TestMathUnitTest::test_floor_1_integer
```
Das Modul [parametrisiert](https://pypi.org/project/parameterized/), das sich bereits in den Entwickler-Abhรคngigkeiten befindet
von `transformers` befindet, funktioniert sowohl fรผr `unittests` als auch fรผr `pytest` Tests.
Wenn es sich bei dem Test jedoch nicht um einen `Unittest` handelt, kรถnnen Sie `pytest.mark.parametrize` verwenden (oder Sie kรถnnen sehen, dass es in
einigen bestehenden Tests verwendet wird, meist unter `Beispiele`).
Hier ist das gleiche Beispiel, diesmal unter Verwendung der `parametrize`-Markierung von `pytest`:
```python
# test_this2.py
import pytest
@pytest.mark.parametrize(
"name, input, expected",
[
("negative", -1.5, -2.0),
("integer", 1, 1.0),
("large fraction", 1.6, 1),
],
)
def test_floor(name, input, expected):
assert_equal(math.floor(input), expected)
```
Genau wie bei `parametrisiert` kรถnnen Sie mit `pytest.mark.parametrize` genau steuern, welche Subtests ausgefรผhrt werden
ausgefรผhrt werden, wenn der Filter `-k` nicht ausreicht. Allerdings erzeugt diese Parametrisierungsfunktion einen etwas anderen Satz von
Namen fรผr die Untertests. Sie sehen folgendermaรen aus:
```bash
pytest test_this2.py --collect-only -q
```
und es wird aufgelistet:
```bash
test_this2.py::test_floor[integer-1-1.0]
test_this2.py::test_floor[negative--1.5--2.0]
test_this2.py::test_floor[large fraction-1.6-1]
```
Jetzt kรถnnen Sie also nur den spezifischen Test durchfรผhren:
```bash
pytest test_this2.py::test_floor[negative--1.5--2.0] test_this2.py::test_floor[integer-1-1.0]
```
wie im vorherigen Beispiel.
### Dateien und Verzeichnisse
In Tests mรผssen wir oft wissen, wo sich Dinge relativ zur aktuellen Testdatei befinden, und das ist nicht trivial, da der Test
von mehreren Verzeichnissen aus aufgerufen werden kann oder sich in Unterverzeichnissen mit unterschiedlicher Tiefe befinden kann. Eine Hilfsklasse
`transformers.test_utils.TestCasePlus` lรถst dieses Problem, indem sie alle grundlegenden Pfade sortiert und einfache
Zugriffsmรถglichkeiten auf sie bietet:
- `pathlib`-Objekte (alle vollstรคndig aufgelรถst):
- `test_file_path` - der aktuelle Testdateipfad, d.h. `__file__`
- `test_file_dir` - das Verzeichnis, das die aktuelle Testdatei enthรคlt
- `tests_dir` - das Verzeichnis der `tests` Testreihe
- `examples_dir` - das Verzeichnis der `examples` Test-Suite
- repo_root_dir` - das Verzeichnis des Repositorys
- src_dir` - das Verzeichnis von `src` (d.h. wo sich das Unterverzeichnis `transformers` befindet)
- stringifizierte Pfade - wie oben, aber diese geben Pfade als Strings zurรผck, anstatt als `pathlib`-Objekte:
- `test_file_path_str`
- `test_file_dir_str`
- `tests_dir_str`
- `examples_dir_str`
- `repo_root_dir_str`
- `src_dir_str`
Um diese zu verwenden, mรผssen Sie lediglich sicherstellen, dass der Test in einer Unterklasse von
`transformers.test_utils.TestCasePlus` befindet. Zum Beispiel:
```python
from transformers.testing_utils import TestCasePlus
class PathExampleTest(TestCasePlus):
def test_something_involving_local_locations(self):
data_dir = self.tests_dir / "fixtures/tests_samples/wmt_en_ro"
```
Wenn Sie Pfade nicht รผber `pathlib` manipulieren mรผssen oder nur einen Pfad als String benรถtigen, kรถnnen Sie jederzeit
`str()` auf das `pathlib`-Objekt anwenden oder die Accessoren mit der Endung `_str` verwenden. Zum Beispiel:
```python
from transformers.testing_utils import TestCasePlus
class PathExampleTest(TestCasePlus):
def test_something_involving_stringified_locations(self):
examples_dir = self.examples_dir_str
```
### Temporรคre Dateien und Verzeichnisse
Die Verwendung eindeutiger temporรคrer Dateien und Verzeichnisse ist fรผr die parallele Durchfรผhrung von Tests unerlรคsslich, damit sich die Tests nicht gegenseitig รผberschreiben.
Daten gegenseitig รผberschreiben. Auรerdem mรถchten wir, dass die temporรคren Dateien und Verzeichnisse am Ende jedes Tests, der sie erstellt hat, gelรถscht werden.
erstellt hat. Daher ist die Verwendung von Paketen wie `tempfile`, die diese Anforderungen erfรผllen, unerlรคsslich.
Beim Debuggen von Tests mรผssen Sie jedoch sehen kรถnnen, was in der temporรคren Datei oder dem temporรคren Verzeichnis gespeichert wird und Sie mรถchten
Sie mรผssen den genauen Pfad kennen und dรผrfen ihn nicht bei jedem neuen Testdurchlauf zufรคllig รคndern.
Fรผr solche Zwecke ist die Hilfsklasse `transformers.test_utils.TestCasePlus` am besten geeignet. Sie ist eine Unterklasse von
Unittest.TestCase`, so dass wir in den Testmodulen einfach von ihr erben kรถnnen.
Hier ist ein Beispiel fรผr die Verwendung dieser Klasse:
```python
from transformers.testing_utils import TestCasePlus
class ExamplesTests(TestCasePlus):
def test_whatever(self):
tmp_dir = self.get_auto_remove_tmp_dir()
```
Dieser Code erstellt ein eindeutiges temporรคres Verzeichnis und setzt `tmp_dir` auf dessen Speicherort.
- Erstellen Sie ein eindeutiges temporรคres Verzeichnis:
```python
def test_whatever(self):
tmp_dir = self.get_auto_remove_tmp_dir()
```
tmp_dir" enthรคlt den Pfad zu dem erstellten temporรคren Verzeichnis. Es wird am Ende des Tests automatisch entfernt.
Tests entfernt.
- Erstellen Sie ein temporรคres Verzeichnis meiner Wahl, stellen Sie sicher, dass es leer ist, bevor der Test beginnt, und leeren Sie es nach dem Test nicht.
```python
def test_whatever(self):
tmp_dir = self.get_auto_remove_tmp_dir("./xxx")
```
Dies ist nรผtzlich fรผr die Fehlersuche, wenn Sie ein bestimmtes Verzeichnis รผberwachen und sicherstellen mรถchten, dass die vorherigen Tests keine Daten darin hinterlassen haben.
keine Daten dort hinterlassen haben.
- Sie kรถnnen das Standardverhalten auรer Kraft setzen, indem Sie die Argumente `before` und `after` direkt รผberschreiben, was zu einem der folgenden Verhaltensweisen fรผhrt
folgenden Verhaltensweisen:
- `before=True`: das temporรคre Verzeichnis wird immer zu Beginn des Tests gelรถscht.
- `before=False`: wenn das temporรคre Verzeichnis bereits existiert, bleiben alle vorhandenen Dateien dort erhalten.
- `after=True`: das temporรคre Verzeichnis wird immer am Ende des Tests gelรถscht.
- `after=False`: das temporรคre Verzeichnis wird am Ende des Tests immer beibehalten.
<Tip>
Um das รquivalent von `rm -r` sicher ausfรผhren zu kรถnnen, sind nur Unterverzeichnisse des Projektarchivs checkout erlaubt, wenn
ein explizites `tmp_dir` verwendet wird, so dass nicht versehentlich ein `/tmp` oder ein รคhnlich wichtiger Teil des Dateisystems vernichtet wird.
d.h. geben Sie bitte immer Pfade an, die mit `./` beginnen.
</Tip>
<Tip>
Jeder Test kann mehrere temporรคre Verzeichnisse registrieren, die alle automatisch entfernt werden, sofern nicht anders gewรผnscht.
anders.
</Tip>
### Temporรคre รberschreibung von sys.path
Wenn Sie `sys.path` vorรผbergehend รผberschreiben mรผssen, um z.B. von einem anderen Test zu importieren, kรถnnen Sie den
Kontextmanager `ExtendSysPath` verwenden. Beispiel:
```python
import os
from transformers.testing_utils import ExtendSysPath
bindir = os.path.abspath(os.path.dirname(__file__))
with ExtendSysPath(f"{bindir}/.."):
from test_trainer import TrainerIntegrationCommon # noqa
```
### รberspringen von Tests
Dies ist nรผtzlich, wenn ein Fehler gefunden und ein neuer Test geschrieben wird, der Fehler aber noch nicht behoben ist. Damit wir ihn
in das Haupt-Repository zu รผbertragen, mรผssen wir sicherstellen, dass er bei `make test` รผbersprungen wird.
Methoden:
- Ein **Skip** bedeutet, dass Sie erwarten, dass Ihr Test nur dann erfolgreich ist, wenn einige Bedingungen erfรผllt sind, andernfalls sollte pytest den Test รผberspringen.
die Ausfรผhrung des Tests ganz รผberspringen. รbliche Beispiele sind das รberspringen von Tests, die nur unter Windows laufen, auf Nicht-Windows-Plattformen oder das รberspringen von
Tests, die von einer externen Ressource abhรคngen, die im Moment nicht verfรผgbar ist (z.B. eine Datenbank).
- Ein **xfail** bedeutet, dass Sie erwarten, dass ein Test aus irgendeinem Grund fehlschlรคgt. Ein gรคngiges Beispiel ist ein Test fรผr eine Funktion, die noch nicht
noch nicht implementiert oder ein noch nicht behobener Fehler. Wenn ein Test trotz eines erwarteten Fehlschlags bestanden wird (markiert mit
pytest.mark.xfail), ist dies ein xpass und wird in der Testzusammenfassung gemeldet.
Einer der wichtigsten Unterschiede zwischen den beiden ist, dass `skip` den Test nicht ausfรผhrt, wรคhrend `xfail` dies tut. Wenn also der
Code, der fehlerhaft ist, einen schlechten Zustand verursacht, der sich auf andere Tests auswirkt, sollten Sie also nicht `xfail` verwenden.
#### Implementierung
- Hier sehen Sie, wie Sie einen ganzen Test bedingungslos รผberspringen kรถnnen:
```python no-style
@unittest.skip("this bug needs to be fixed")
def test_feature_x():
```
oder mit pytest:
```python no-style
@pytest.mark.skip(reason="this bug needs to be fixed")
```
oder mit dem `xfail` Weg:
```python no-style
@pytest.mark.xfail
def test_feature_x():
```
- Hier erfahren Sie, wie Sie einen Test aufgrund einer internen Prรผfung innerhalb des Tests auslassen kรถnnen:
```python
def test_feature_x():
if not has_something():
pytest.skip("unsupported configuration")
```
oder das ganze Modul:
```python
import pytest
if not pytest.config.getoption("--custom-flag"):
pytest.skip("--custom-flag is missing, skipping tests", allow_module_level=True)
```
oder mit dem `xfail` Weg:
```python
def test_feature_x():
pytest.xfail("expected to fail until bug XYZ is fixed")
```
- Hier erfahren Sie, wie Sie alle Tests in einem Modul รผberspringen kรถnnen, wenn ein Import fehlt:
```python
docutils = pytest.importorskip("docutils", minversion="0.3")
```
- Einen Test aufgrund einer Bedingung รผberspringen:
```python no-style
@pytest.mark.skipif(sys.version_info < (3,6), reason="requires python3.6 or higher")
def test_feature_x():
```
oder:
```python no-style
@unittest.skipIf(torch_device == "cpu", "Can't do half precision")
def test_feature_x():
```
oder รผberspringen Sie das ganze Modul:
```python no-style
@pytest.mark.skipif(sys.platform == 'win32', reason="does not run on windows")
class TestClass():
def test_feature_x(self):
```
Weitere Details, Beispiele und Mรถglichkeiten finden Sie [hier](https://docs.pytest.org/en/latest/skipping.html).
### Langsame Tests
Die Bibliothek der Tests wรคchst stรคndig, und einige der Tests brauchen Minuten, um ausgefรผhrt zu werden, daher kรถnnen wir es uns nicht leisten, eine Stunde zu warten, bis die
eine Stunde auf die Fertigstellung der Testsuite auf CI zu warten. Daher sollten langsame Tests, mit einigen Ausnahmen fรผr wichtige Tests, wie im folgenden Beispiel
wie im folgenden Beispiel markiert werden:
```python no-style
from transformers.testing_utils import slow
@slow
def test_integration_foo():
```
Sobald ein Test als `@langsam` markiert ist, setzen Sie die Umgebungsvariable `RUN_SLOW=1`, um solche Tests auszufรผhren, z.B:
```bash
RUN_SLOW=1 pytest tests
```
Einige Dekoratoren wie `@parameterized` schreiben Testnamen um, daher mรผssen `@slow` und die รผbrigen Skip-Dekoratoren
`@require_*` mรผssen als letztes aufgefรผhrt werden, damit sie korrekt funktionieren. Hier ist ein Beispiel fรผr die korrekte Verwendung:
```python no-style
@parameteriz ed.expand(...)
@slow
def test_integration_foo():
```
Wie zu Beginn dieses Dokuments erlรคutert, werden langsame Tests nach einem Zeitplan ausgefรผhrt und nicht in PRs CI
Prรผfungen. Es ist also mรถglich, dass einige Probleme bei der Einreichung eines PRs รผbersehen werden und zusammengefรผhrt werden. Solche Probleme werden
werden beim nรคchsten geplanten CI-Job abgefangen. Das bedeutet aber auch, dass es wichtig ist, die langsamen Tests auf Ihrem
Rechner auszufรผhren, bevor Sie den PR einreichen.
Hier ist ein grober Entscheidungsmechanismus fรผr die Auswahl der Tests, die als langsam markiert werden sollen:
Wenn der Test auf eine der internen Komponenten der Bibliothek ausgerichtet ist (z.B. Modellierungsdateien, Tokenisierungsdateien,
Pipelines), dann sollten wir diesen Test in der nicht langsamen Testsuite ausfรผhren. Wenn er sich auf einen anderen Aspekt der Bibliothek bezieht,
wie z.B. die Dokumentation oder die Beispiele, dann sollten wir diese Tests in der langsamen Testsuite durchfรผhren. Und dann, zur Verfeinerung
Ansatz zu verfeinern, sollten wir Ausnahmen einfรผhren:
- Alle Tests, die einen umfangreichen Satz von Gewichten oder einen Datensatz mit einer Grรถรe von mehr als ~50MB herunterladen mรผssen (z.B. Modell- oder
Tokenizer-Integrationstests, Pipeline-Integrationstests) sollten auf langsam gesetzt werden. Wenn Sie ein neues Modell hinzufรผgen, sollten Sie
sollten Sie eine kleine Version des Modells (mit zufรคlligen Gewichtungen) fรผr Integrationstests erstellen und in den Hub hochladen. Dies wird
wird in den folgenden Abschnitten erlรคutert.
- Alle Tests, die ein Training durchfรผhren mรผssen, das nicht speziell auf Schnelligkeit optimiert ist, sollten auf langsam gesetzt werden.
- Wir kรถnnen Ausnahmen einfรผhren, wenn einige dieser Tests, die nicht langsam sein sollten, unertrรคglich langsam sind, und sie auf
@langsam`. Auto-Modellierungstests, die groรe Dateien auf der Festplatte speichern und laden, sind ein gutes Beispiel fรผr Tests, die als
als `@langsam` markiert sind.
- Wenn ein Test in weniger als 1 Sekunde auf CI abgeschlossen wird (einschlieรlich eventueller Downloads), sollte es sich trotzdem um einen normalen Test handeln.
Insgesamt mรผssen alle nicht langsamen Tests die verschiedenen Interna abdecken und dabei schnell bleiben. Zum Beispiel,
kann eine signifikante Abdeckung erreicht werden, indem Sie mit speziell erstellten kleinen Modellen mit zufรคlligen Gewichten testen. Solche Modelle
haben eine sehr geringe Anzahl von Schichten (z.B. 2), Vokabeln (z.B. 1000), usw. Dann kรถnnen die `@slow`-Tests groรe
langsame Modelle verwenden, um qualitative Tests durchzufรผhren. Um die Verwendung dieser Modelle zu sehen, suchen Sie einfach nach *winzigen* Modellen mit:
```bash
grep tiny tests examples
```
Hier ist ein Beispiel fรผr ein [Skript](https://github.com/huggingface/transformers/tree/main/scripts/fsmt/fsmt-make-tiny-model.py), das das winzige Modell erstellt hat
[stas/tiny-wmt19-en-de](https://huggingface.co/stas/tiny-wmt19-en-de). Sie kรถnnen es ganz einfach an Ihre eigene
Architektur Ihres Modells anpassen.
Es ist leicht, die Laufzeit falsch zu messen, wenn zum Beispiel ein groรes Modell heruntergeladen wird, aber wenn
Sie es lokal testen, wรผrden die heruntergeladenen Dateien zwischengespeichert und somit die Download-Zeit nicht gemessen werden. Prรผfen Sie daher den
Ausfรผhrungsgeschwindigkeitsbericht in den CI-Protokollen (die Ausgabe von `pytest --durations=0 tests`).
Dieser Bericht ist auch nรผtzlich, um langsame Ausreiรer zu finden, die nicht als solche gekennzeichnet sind oder die neu geschrieben werden mรผssen, um schnell zu sein.
Wenn Sie bemerken, dass die Testsuite beim CI langsam wird, zeigt die oberste Liste dieses Berichts die langsamsten
Tests.
### Testen der stdout/stderr-Ausgabe
Um Funktionen zu testen, die in `stdout` und/oder `stderr` schreiben, kann der Test auf diese Strรถme zugreifen, indem er die
[capsys system](https://docs.pytest.org/en/latest/capture.html) von `pytest` zugreifen. So wird dies bewerkstelligt:
```python
import sys
def print_to_stdout(s):
print(s)
def print_to_stderr(s):
sys.stderr.write(s)
def test_result_and_stdout(capsys):
msg = "Hello"
print_to_stdout(msg)
print_to_stderr(msg)
out, err = capsys.readouterr() # consume the captured output streams
# optional: if you want to replay the consumed streams:
sys.stdout.write(out)
sys.stderr.write(err)
# test:
assert msg in out
assert msg in err
```
Und natรผrlich wird `stderr` in den meisten Fรคllen als Teil einer Ausnahme auftreten, so dass try/except in einem solchen Fall verwendet werden muss
Fall verwendet werden:
```python
def raise_exception(msg):
raise ValueError(msg)
def test_something_exception():
msg = "Not a good value"
error = ""
try:
raise_exception(msg)
except Exception as e:
error = str(e)
assert msg in error, f"{msg} is in the exception:\n{error}"
```
Ein anderer Ansatz zur Erfassung von stdout ist `contextlib.redirect_stdout`:
```python
from io import StringIO
from contextlib import redirect_stdout
def print_to_stdout(s):
print(s)
def test_result_and_stdout():
msg = "Hello"
buffer = StringIO()
with redirect_stdout(buffer):
print_to_stdout(msg)
out = buffer.getvalue()
# optional: if you want to replay the consumed streams:
sys.stdout.write(out)
# test:
assert msg in out
```
Ein wichtiges potenzielles Problem beim Erfassen von stdout ist, dass es `r` Zeichen enthalten kann, die bei normalem `print`
alles zurรผcksetzen, was bisher gedruckt wurde. Mit `pytest` gibt es kein Problem, aber mit `pytest -s` werden diese
werden diese Zeichen in den Puffer aufgenommen. Um den Test mit und ohne `-s` laufen zu lassen, mรผssen Sie also eine zusรคtzliche Bereinigung
zusรคtzliche Bereinigung der erfassten Ausgabe vornehmen, indem Sie `re.sub(r'~.*\r', '', buf, 0, re.M)` verwenden.
Aber dann haben wir einen Hilfskontextmanager-Wrapper, der sich automatisch um alles kรผmmert, unabhรคngig davon, ob er
einige "*.*.*.*" enthรคlt oder nicht:
```python
from transformers.testing_utils import CaptureStdout
with CaptureStdout() as cs:
function_that_writes_to_stdout()
print(cs.out)
```
Hier ist ein vollstรคndiges Testbeispiel:
```python
from transformers.testing_utils import CaptureStdout
msg = "Secret message\r"
final = "Hello World"
with CaptureStdout() as cs:
print(msg + final)
assert cs.out == final + "\n", f"captured: {cs.out}, expecting {final}"
```
Wenn Sie `stderr` aufzeichnen mรถchten, verwenden Sie stattdessen die Klasse `CaptureStderr`:
```python
from transformers.testing_utils import CaptureStderr
with CaptureStderr() as cs:
function_that_writes_to_stderr()
print(cs.err)
```
Wenn Sie beide Streams auf einmal erfassen mรผssen, verwenden Sie die รผbergeordnete Klasse `CaptureStd`:
```python
from transformers.testing_utils import CaptureStd
with CaptureStd() as cs:
function_that_writes_to_stdout_and_stderr()
print(cs.err, cs.out)
```
Um das Debuggen von Testproblemen zu erleichtern, geben diese Kontextmanager standardmรครig die aufgezeichneten Streams beim Verlassen
aus dem Kontext wieder.
### Erfassen von Logger-Streams
Wenn Sie die Ausgabe eines Loggers validieren mรผssen, kรถnnen Sie `CaptureLogger` verwenden:
```python
from transformers import logging
from transformers.testing_utils import CaptureLogger
msg = "Testing 1, 2, 3"
logging.set_verbosity_info()
logger = logging.get_logger("transformers.models.bart.tokenization_bart")
with CaptureLogger(logger) as cl:
logger.info(msg)
assert cl.out, msg + "\n"
```
### Testen mit Umgebungsvariablen
Wenn Sie die Auswirkungen von Umgebungsvariablen fรผr einen bestimmten Test testen mรถchten, kรถnnen Sie einen Hilfsdekorator verwenden
`transformers.testing_utils.mockenv`
```python
from transformers.testing_utils import mockenv
class HfArgumentParserTest(unittest.TestCase):
@mockenv(TRANSFORMERS_VERBOSITY="error")
def test_env_override(self):
env_level_str = os.getenv("TRANSFORMERS_VERBOSITY", None)
```
Manchmal muss ein externes Programm aufgerufen werden, was die Einstellung von `PYTHONPATH` in `os.environ` erfordert, um mehrere lokale Pfade einzuschlieรen.
mehrere lokale Pfade. Eine Hilfsklasse `transformers.test_utils.TestCasePlus` hilft Ihnen dabei:
```python
from transformers.testing_utils import TestCasePlus
class EnvExampleTest(TestCasePlus):
def test_external_prog(self):
env = self.get_env()
# now call the external program, passing `env` to it
```
Je nachdem, ob die Testdatei in der Testsuite `tests` oder in `examples` war, wird sie korrekt eingerichtet
env[PYTHONPATH]` eines dieser beiden Verzeichnisse und auch das `src` Verzeichnis, um sicherzustellen, dass der Test gegen das aktuelle
um sicherzustellen, dass der Test mit dem aktuellen Projektarchiv durchgefรผhrt wird, und schlieรlich mit dem, was in `env[PYTHONPATH]` bereits eingestellt war, bevor der Test aufgerufen wurde.
wenn รผberhaupt.
Diese Hilfsmethode erstellt eine Kopie des Objekts `os.environ`, so dass das Original intakt bleibt.
### Reproduzierbare Ergebnisse erhalten
In manchen Situationen mรถchten Sie vielleicht die Zufรคlligkeit Ihrer Tests beseitigen. Um identische, reproduzierbare Ergebnisse zu erhalten, mรผssen Sie
mรผssen Sie den Seed festlegen:
```python
seed = 42
# python RNG
import random
random.seed(seed)
# pytorch RNGs
import torch
torch.manual_seed(seed)
torch.backends.cudnn.deterministic = True
if torch.cuda.is_available():
torch.cuda.manual_seed_all(seed)
# numpy RNG
import numpy as np
np.random.seed(seed)
# tf RNG
tf.random.set_seed(seed)
```
### Tests debuggen
Um einen Debugger an der Stelle zu starten, an der die Warnung auftritt, gehen Sie wie folgt vor:
```bash
pytest tests/utils/test_logging.py -W error::UserWarning --pdb
```
## Arbeiten mit Github-Aktionen-Workflows
Um einen CI-Job fรผr einen Self-Push-Workflow auszulรถsen, mรผssen Sie:
1. Erstellen Sie einen neuen Zweig auf `transformers` Ursprung (keine Gabelung!).
2. Der Name der Verzweigung muss entweder mit `ci_` oder `ci-` beginnen (`main` lรถst ihn auch aus, aber wir kรถnnen keine PRs auf
`main`). Es wird auch nur fรผr bestimmte Pfade ausgelรถst - Sie kรถnnen die aktuelle Definition finden, falls sie
falls sie sich seit der Erstellung dieses Dokuments geรคndert hat [hier](https://github.com/huggingface/transformers/blob/main/.github/workflows/self-push.yml) unter *push:*
3. Erstellen Sie einen PR von diesem Zweig.
4. Dann kรถnnen Sie sehen, wie der Job erscheint [hier](https://github.com/huggingface/transformers/actions/workflows/self-push.yml). Er wird mรถglicherweise nicht sofort ausgefรผhrt, wenn es
ein Backlog vorhanden ist.
## Testen experimenteller CI-Funktionen
Das Testen von CI-Funktionen kann potenziell problematisch sein, da es die normale CI-Funktion beeintrรคchtigen kann. Wenn also eine
neue CI-Funktion hinzugefรผgt werden soll, sollte dies wie folgt geschehen.
1. Erstellen Sie einen neuen Auftrag, der die zu testende Funktion testet.
2. Der neue Job muss immer erfolgreich sein, so dass er uns ein grรผnes โ gibt (Details unten).
3. Lassen Sie ihn einige Tage lang laufen, um zu sehen, dass eine Vielzahl verschiedener PR-Typen darauf laufen (Benutzer-Gabelzweige,
nicht geforkte Zweige, Zweige, die von github.com UI direct file edit stammen, verschiedene erzwungene Pushes, etc. - es gibt
es gibt so viele), wรคhrend Sie die Protokolle des experimentellen Jobs รผberwachen (nicht den gesamten Job grรผn, da er absichtlich immer
grรผn)
4. Wenn klar ist, dass alles in Ordnung ist, fรผgen Sie die neuen รnderungen in die bestehenden Jobs ein.
Auf diese Weise wird der normale Arbeitsablauf nicht durch Experimente mit der CI-Funktionalitรคt selbst beeintrรคchtigt.
Wie kรถnnen wir nun dafรผr sorgen, dass der Auftrag immer erfolgreich ist, wรคhrend die neue CI-Funktion entwickelt wird?
Einige CIs, wie TravisCI, unterstรผtzen ignore-step-failure und melden den gesamten Job als erfolgreich, aber CircleCI und
Github Actions unterstรผtzen dies zum jetzigen Zeitpunkt nicht.
Sie kรถnnen also die folgende Abhilfe verwenden:
1. Setzen Sie `set +euo pipefail` am Anfang des Ausfรผhrungsbefehls, um die meisten potenziellen Fehler im Bash-Skript zu unterdrรผcken.
2. Der letzte Befehl muss ein Erfolg sein: `echo "done"` oder einfach `true` reicht aus.
Hier ist ein Beispiel:
```yaml
- run:
name: run CI experiment
command: |
set +euo pipefail
echo "setting run-all-despite-any-errors-mode"
this_command_will_fail
echo "but bash continues to run"
# emulate another failure
false
# but the last command must be a success
echo "during experiment do not remove: reporting success to CI, even if there were failures"
```
Fรผr einfache Befehle kรถnnen Sie auch Folgendes tun:
```bash
cmd_that_may_fail || true
```
Wenn Sie mit den Ergebnissen zufrieden sind, integrieren Sie den experimentellen Schritt oder Job natรผrlich in den Rest der normalen Jobs,
Entfernen Sie dabei `set +euo pipefail` oder andere Dinge, die Sie eventuell hinzugefรผgt haben, um sicherzustellen, dass der experimentelle Auftrag nicht
den normalen CI-Betrieb nicht beeintrรคchtigt.
Dieser ganze Prozess wรคre viel einfacher gewesen, wenn wir nur etwas wie `allow-failure` fรผr den
experimentellen Schritt festlegen kรถnnten und ihn scheitern lassen wรผrden, ohne den Gesamtstatus der PRs zu beeintrรคchtigen. Aber wie bereits erwรคhnt, haben CircleCI und
Github Actions dies im Moment nicht unterstรผtzen.
Sie kรถnnen in diesen CI-spezifischen Threads fรผr diese Funktion stimmen und sehen, wo sie steht:
- [Github Actions:](https://github.com/actions/toolkit/issues/399)
- [CircleCI:](https://ideas.circleci.com/ideas/CCI-I-344)
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/fr/in_translation.md
|
<!--โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Traduction en cours.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/fr/_toctree.yml
|
- sections:
- local: index
title: ๐ค Transformers
- local: quicktour
title: Visite rapide
- local: in_translation
title: Installation
title: Dรฉmarrer
- sections:
- local: in_translation
title: Pipelines pour l'infรฉrence
- local: in_translation
title: Chargement d'instances prรฉ-entraรฎnรฉes avec une AutoClass
- local: in_translation
title: Prรฉparation des donnรฉes
- local: in_translation
title: Fine-tune un modรจle prรฉ-entraรฎnรฉ
- local: in_translation
title: Entraรฎnement distribuรฉ avec ๐ค Accelerate
- local: in_translation
title: Partager un modรจle
title: Tutoriels
- sections:
- sections:
- local: in_translation
title: Crรฉer votre architecture
- local: in_translation
title: Partager vos modรจles
- local: in_translation
title: Entraรฎnement avec un script
- local: in_translation
title: Entraรฎnement avec Amazon SageMaker
- local: in_translation
title: Convertir depuis des checkpoints Tensorflow
- local: in_translation
title: Exporter vers ONNX
- local: in_translation
title: Exporter vers TorchScript
- local: in_translation
title: Aide au dรฉpannage
title: Usage gรฉnรฉral
- sections:
- local: in_translation
title: Utiliser les tokenizers de ๐ค Tokenizers
- local: in_translation
title: Infรฉrence avec les modรจles multilingues
- local: in_translation
title: Stratรฉgies de gรฉnรฉration de texte
- sections:
- isExpanded: false
local: in_translation
title: Classification de texte
- local: in_translation
title: Classification de token
- local: in_translation
title: Systรจme de question-rรฉponse
- local: in_translation
title: Modรฉlisation causale du langage
- local: in_translation
title: Modรฉlisation du langage avec masque
- local: in_translation
title: Traduction
- local: in_translation
title: Gรฉnรฉration de rรฉsumรฉ
- local: in_translation
title: Question ร choix multiple
title: Guides des tรขches
title: Traitement automatique des langues
- sections:
- local: in_translation
title: Classification audio
- local: in_translation
title: Reconnaissance automatique de la parole
title: Audio
- sections:
- local: in_translation
title: Classification d'images
- local: in_translation
title: Segmentation sรฉmantique
- local: in_translation
title: Classification de vidรฉos
- local: in_translation
title: Dรฉtection d'objets
title: Vision par ordinateur
- sections:
- local: in_translation
title: Performance et extensibilitรฉ
- sections:
- local: in_translation
title: Comment contribuer ร transformers?
- local: in_translation
title: Comment ajouter un modรจle ร ๐ค Transformers?
- local: in_translation
title: Comment convertir un modรจle ๐ค Transformers vers TensorFlow?
- local: in_translation
title: Comment ajouter un pipeline ร ๐ค Transformers?
- local: in_translation
title: Tester
- local: in_translation
title: Vรฉrification pour une Pull Request
title: Contribuer
- local: in_translation
title: ๐ค Transformers Notebooks
- local: in_translation
title: Ressources communautaires
- local: in_translation
title: Benchmarks
- local: in_translation
title: Migration ร partir de versions prรฉcรฉdentes
title: Guides d'utilisation
- sections:
- local: in_translation
title: Philosophie
- local: in_translation
title: Glossaire
- local: in_translation
title: Qu'est ce ๐ค Transformers peut faire ?
- local: in_translation
title: Quelles tรขches ๐ค Transformers peut rรฉsoudre ?
- local: in_translation
title: Rรฉsumรฉ des modรจles
- local: in_translation
title: Rรฉsumรฉ des tokenizers
- local: in_translation
title: Remplissage et troncature
- local: in_translation
title: BERTology
- local: in_translation
title: Perplexitรฉ des modรจles ร longueur fixe
- local: in_translation
title: Pipelines pour infรฉrence avec des serveurs web
title: Guides conceptuels
- sections:
- isExpanded: false
sections:
- local: in_translation
title: Classes principales
- local: in_translation
title: Modรจles textuels
- local: in_translation
title: Modรจles visuels
- local: in_translation
title: Modรจles audio
- local: in_translation
title: Modรจles multimodal
- local: in_translation
title: Modรจles d'apprentissage par renforcement
- local: in_translation
title: Modรจles de sรฉries temporelles
- local: in_translation
title: Graph models
title: Modรจles
- sections:
- local: in_translation
title: Utilitaires internes
title: API
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/fr/_config.py
|
# docstyle-ignore
INSTALL_CONTENT = """
# Installation de Transformers
! pip install transformers datasets
# Pour installer ร partir du code source au lieu de la derniรจre version, commentez la commande ci-dessus et dรฉcommentez la suivante.
# ! pip install git+https://github.com/huggingface/transformers.git
"""
notebook_first_cells = [{"type": "code", "content": INSTALL_CONTENT}]
black_avoid_patterns = {
"{processor_class}": "FakeProcessorClass",
"{model_class}": "FakeModelClass",
"{object_class}": "FakeObjectClass",
}
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/fr/index.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ๐ค Transformers
Apprentissage automatique de pointe pour [PyTorch](https://pytorch.org/), [TensorFlow](https://www.tensorflow.org/), et [JAX](https://jax.readthedocs.io/en/latest/).
๐ค Transformers fournit des API et des outils pour tรฉlรฉcharger et entraรฎner facilement des modรจles prรฉ-entraรฎnรฉs de pointe. L'utilisation de modรจles prรฉ-entraรฎnรฉs peut rรฉduire vos coรปts de calcul, votre empreinte carbone, et vous faire รฉconomiser le temps et les ressources nรฉcessaires pour entraรฎner un modรจle ร partir de zรฉro. Ces modรจles prennent en charge des tรขches courantes dans diffรฉrentes modalitรฉs, telles que :
๐ **Traitement automatique des langues**: classification de texte, reconnaissance d'entitรฉs, systรจme de question-rรฉponse, modรจle de langage, gรฉnรฉration de rรฉsumรฉ, traduction, question ร choix multiples et gรฉnรฉration de texte.<br>
๐ผ๏ธ **Vision par ordinateur**: classification d'image, dรฉtection d'objet et segmentation.<br>
๐ฃ๏ธ **Audio**: reconnaissance automatique de la parole et classification audio.<br>
๐ **Multimodalitรฉ**: systรจme de question-rรฉponse avec des tableaux ou images, reconnaissance optique de caractรจres, extraction d'information depuis des documents scannรฉs et classification de vidรฉo.
๐ค Transformers prend en charge l'interopรฉrabilitรฉ entre PyTorch, TensorFlow et JAX. Cela permet d'utiliser un framework diffรฉrent ร chaque รฉtape de la vie d'un modรจle, par exemple entraรฎner un modรจle en trois lignes de code avec un framework, et le charger pour l'infรฉrence avec un autre. Les modรจles peuvent รฉgalement รชtre exportรฉs dans un format comme ONNX et TorchScript pour รชtre dรฉployรฉs dans des environnements de production.
Rejoignez la communautรฉ grandissante sur le [Hub](https://huggingface.co/models), le [forum](https://discuss.huggingface.co/) ou [Discord](https://discord.com/invite/JfAtkvEtRb) dรจs aujourd'hui !
## Si vous cherchez un support personnalisรฉ de l'รฉquipe Hugging Face
<a target="_blank" href="https://huggingface.co/support">
<img alt="HuggingFace Expert Acceleration Program" src="https://cdn-media.huggingface.co/marketing/transformers/new-support-improved.png" style="width: 100%; max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
</a>
## Contents
La documentation est organisรฉe en 5 parties:
- **DEMARRER** propose une visite rapide de la bibliothรจque et des instructions d'installation pour รชtre opรฉrationnel.
- **TUTORIELS** excellent point de dรฉpart pour les dรฉbutants. Cette section vous aidera ร acquรฉrir les compรฉtences de base dont vous avez besoin pour commencer ร utiliser la bibliothรจque.
- **GUIDES D'UTILISATION** pour diffรฉrentes tรขches comme par exemple le finetuning d'un modรจle prรฉ-entraรฎnรฉ pour la classification de texte ou comment crรฉer et partager votre propre modรจle.
- **GUIDES CONCEPTUELS** pour plus de discussions et d'explications sur les concepts et les idรฉes sous-jacentes aux modรจles, aux tรขches et ร la philosophie de conception de ๐ค Transformers.
- **API** dรฉcrit toutes les classes et fonctions :
- **CLASSES PRINCIPALES** dรฉtaille les classes les plus importantes comme la configuration, le modรจle, le tokenizer et le pipeline..
- **MODELES** dรฉtaille les classes et les fonctions propres ร chaque modรจle de la bibliothรจque.
- **UTILITAIRES INTERNES** dรฉtaille les classes et fonctions utilitaires utilisรฉes en interne.
### Modรจles supportรฉs
<!--This list is updated automatically from the README with _make fix-copies_. Do not update manually! -->
1. **[ALBERT](model_doc/albert)** (from Google Research and the Toyota Technological Institute at Chicago) released with the paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), by Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
1. **[ALIGN](model_doc/align)** (from Google Research) released with the paper [Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision](https://arxiv.org/abs/2102.05918) by Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc V. Le, Yunhsuan Sung, Zhen Li, Tom Duerig.
1. **[AltCLIP](model_doc/altclip)** (from BAAI) released with the paper [AltCLIP: Altering the Language Encoder in CLIP for Extended Language Capabilities](https://arxiv.org/abs/2211.06679) by Chen, Zhongzhi and Liu, Guang and Zhang, Bo-Wen and Ye, Fulong and Yang, Qinghong and Wu, Ledell.
1. **[Audio Spectrogram Transformer](model_doc/audio-spectrogram-transformer)** (from MIT) released with the paper [AST: Audio Spectrogram Transformer](https://arxiv.org/abs/2104.01778) by Yuan Gong, Yu-An Chung, James Glass.
1. **[BART](model_doc/bart)** (from Facebook) released with the paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) by Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov and Luke Zettlemoyer.
1. **[BARThez](model_doc/barthez)** (from รcole polytechnique) released with the paper [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) by Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis.
1. **[BARTpho](model_doc/bartpho)** (from VinAI Research) released with the paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) by Nguyen Luong Tran, Duong Minh Le and Dat Quoc Nguyen.
1. **[BEiT](model_doc/beit)** (from Microsoft) released with the paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) by Hangbo Bao, Li Dong, Furu Wei.
1. **[BERT](model_doc/bert)** (from Google) released with the paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) by Jacob Devlin, Ming-Wei Chang, Kenton Lee and Kristina Toutanova.
1. **[BERT For Sequence Generation](model_doc/bert-generation)** (from Google) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
1. **[BERTweet](model_doc/bertweet)** (from VinAI Research) released with the paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) by Dat Quoc Nguyen, Thanh Vu and Anh Tuan Nguyen.
1. **[BigBird-Pegasus](model_doc/bigbird_pegasus)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
1. **[BigBird-RoBERTa](model_doc/big_bird)** (from Google Research) released with the paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) by Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
1. **[BioGpt](model_doc/biogpt)** (from Microsoft Research AI4Science) released with the paper [BioGPT: generative pre-trained transformer for biomedical text generation and mining](https://academic.oup.com/bib/advance-article/doi/10.1093/bib/bbac409/6713511?guestAccessKey=a66d9b5d-4f83-4017-bb52-405815c907b9) by Renqian Luo, Liai Sun, Yingce Xia, Tao Qin, Sheng Zhang, Hoifung Poon and Tie-Yan Liu.
1. **[BiT](model_doc/bit)** (from Google AI) released with the paper [Big Transfer (BiT): General Visual Representation Learning](https://arxiv.org/abs/1912.11370) by Alexander Kolesnikov, Lucas Beyer, Xiaohua Zhai, Joan Puigcerver, Jessica Yung, Sylvain Gelly, Neil Houlsby.
1. **[Blenderbot](model_doc/blenderbot)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
1. **[BlenderbotSmall](model_doc/blenderbot-small)** (from Facebook) released with the paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) by Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
1. **[BLIP](model_doc/blip)** (from Salesforce) released with the paper [BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation](https://arxiv.org/abs/2201.12086) by Junnan Li, Dongxu Li, Caiming Xiong, Steven Hoi.
1. **[BLOOM](model_doc/bloom)** (from BigScience workshop) released by the [BigScience Workshop](https://bigscience.huggingface.co/).
1. **[BORT](model_doc/bort)** (from Alexa) released with the paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) by Adrian de Wynter and Daniel J. Perry.
1. **[BridgeTower](model_doc/bridgetower)** (from Harbin Institute of Technology/Microsoft Research Asia/Intel Labs) released with the paper [BridgeTower: Building Bridges Between Encoders in Vision-Language Representation Learning](https://arxiv.org/abs/2206.08657) by Xiao Xu, Chenfei Wu, Shachar Rosenman, Vasudev Lal, Wanxiang Che, Nan Duan.
1. **[ByT5](model_doc/byt5)** (from Google Research) released with the paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) by Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
1. **[CamemBERT](model_doc/camembert)** (from Inria/Facebook/Sorbonne) released with the paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) by Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suรกrez*, Yoann Dupont, Laurent Romary, รric Villemonte de la Clergerie, Djamรฉ Seddah and Benoรฎt Sagot.
1. **[CANINE](model_doc/canine)** (from Google Research) released with the paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) by Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
1. **[Chinese-CLIP](model_doc/chinese_clip)** (from OFA-Sys) released with the paper [Chinese CLIP: Contrastive Vision-Language Pretraining in Chinese](https://arxiv.org/abs/2211.01335) by An Yang, Junshu Pan, Junyang Lin, Rui Men, Yichang Zhang, Jingren Zhou, Chang Zhou.
1. **[CLIP](model_doc/clip)** (from OpenAI) released with the paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) by Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
1. **[CLIPSeg](model_doc/clipseg)** (from University of Gรถttingen) released with the paper [Image Segmentation Using Text and Image Prompts](https://arxiv.org/abs/2112.10003) by Timo Lรผddecke and Alexander Ecker.
1. **[CodeGen](model_doc/codegen)** (from Salesforce) released with the paper [A Conversational Paradigm for Program Synthesis](https://arxiv.org/abs/2203.13474) by Erik Nijkamp, Bo Pang, Hiroaki Hayashi, Lifu Tu, Huan Wang, Yingbo Zhou, Silvio Savarese, Caiming Xiong.
1. **[Conditional DETR](model_doc/conditional_detr)** (from Microsoft Research Asia) released with the paper [Conditional DETR for Fast Training Convergence](https://arxiv.org/abs/2108.06152) by Depu Meng, Xiaokang Chen, Zejia Fan, Gang Zeng, Houqiang Li, Yuhui Yuan, Lei Sun, Jingdong Wang.
1. **[ConvBERT](model_doc/convbert)** (from YituTech) released with the paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) by Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
1. **[ConvNeXT](model_doc/convnext)** (from Facebook AI) released with the paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) by Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
1. **[ConvNeXTV2](model_doc/convnextv2)** (from Facebook AI) released with the paper [ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders](https://arxiv.org/abs/2301.00808) by Sanghyun Woo, Shoubhik Debnath, Ronghang Hu, Xinlei Chen, Zhuang Liu, In So Kweon, Saining Xie.
1. **[CPM](model_doc/cpm)** (from Tsinghua University) released with the paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) by Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
1. **[CTRL](model_doc/ctrl)** (from Salesforce) released with the paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) by Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong and Richard Socher.
1. **[CvT](model_doc/cvt)** (from Microsoft) released with the paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) by Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
1. **[Data2Vec](model_doc/data2vec)** (from Facebook) released with the paper [Data2Vec: A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) by Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
1. **[DeBERTa](model_doc/deberta)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
1. **[DeBERTa-v2](model_doc/deberta-v2)** (from Microsoft) released with the paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) by Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
1. **[Decision Transformer](model_doc/decision_transformer)** (from Berkeley/Facebook/Google) released with the paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) by Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
1. **[Deformable DETR](model_doc/deformable_detr)** (from SenseTime Research) released with the paper [Deformable DETR: Deformable Transformers for End-to-End Object Detection](https://arxiv.org/abs/2010.04159) by Xizhou Zhu, Weijie Su, Lewei Lu, Bin Li, Xiaogang Wang, Jifeng Dai.
1. **[DeiT](model_doc/deit)** (from Facebook) released with the paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) by Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervรฉ Jรฉgou.
1. **[DETA](model_doc/deta)** (from The University of Texas at Austin) released with the paper [NMS Strikes Back](https://arxiv.org/abs/2212.06137) by Jeffrey Ouyang-Zhang, Jang Hyun Cho, Xingyi Zhou, Philipp Krรคhenbรผhl.
1. **[DETR](model_doc/detr)** (from Facebook) released with the paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) by Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
1. **[DialoGPT](model_doc/dialogpt)** (from Microsoft Research) released with the paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) by Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
1. **[DiNAT](model_doc/dinat)** (from SHI Labs) released with the paper [Dilated Neighborhood Attention Transformer](https://arxiv.org/abs/2209.15001) by Ali Hassani and Humphrey Shi.
1. **[DistilBERT](model_doc/distilbert)** (from HuggingFace), released together with the paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) by Victor Sanh, Lysandre Debut and Thomas Wolf. The same method has been applied to compress GPT2 into [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), RoBERTa into [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), Multilingual BERT into [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation) and a German version of DistilBERT.
1. **[DiT](model_doc/dit)** (from Microsoft Research) released with the paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) by Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
1. **[Donut](model_doc/donut)** (from NAVER), released together with the paper [OCR-free Document Understanding Transformer](https://arxiv.org/abs/2111.15664) by Geewook Kim, Teakgyu Hong, Moonbin Yim, Jeongyeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, Seunghyun Park.
1. **[DPR](model_doc/dpr)** (from Facebook) released with the paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) by Vladimir Karpukhin, Barlas Oฤuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, and Wen-tau Yih.
1. **[DPT](master/model_doc/dpt)** (from Intel Labs) released with the paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) by Renรฉ Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
1. **[EfficientFormer](model_doc/efficientformer)** (from Snap Research) released with the paper [EfficientFormer: Vision Transformers at MobileNetSpeed](https://arxiv.org/abs/2206.01191) by Yanyu Li, Geng Yuan, Yang Wen, Ju Hu, Georgios Evangelidis, Sergey Tulyakov, Yanzhi Wang, Jian Ren.
1. **[ELECTRA](model_doc/electra)** (from Google Research/Stanford University) released with the paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) by Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
1. **[EncoderDecoder](model_doc/encoder-decoder)** (from Google Research) released with the paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) by Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
1. **[ERNIE](model_doc/ernie)** (from Baidu) released with the paper [ERNIE: Enhanced Representation through Knowledge Integration](https://arxiv.org/abs/1904.09223) by Yu Sun, Shuohuan Wang, Yukun Li, Shikun Feng, Xuyi Chen, Han Zhang, Xin Tian, Danxiang Zhu, Hao Tian, Hua Wu.
1. **[ESM](model_doc/esm)** (from Meta AI) are transformer protein language models. **ESM-1b** was released with the paper [Biological structure and function emerge from scaling unsupervised learning to 250 million protein sequences](https://www.pnas.org/content/118/15/e2016239118) by Alexander Rives, Joshua Meier, Tom Sercu, Siddharth Goyal, Zeming Lin, Jason Liu, Demi Guo, Myle Ott, C. Lawrence Zitnick, Jerry Ma, and Rob Fergus. **ESM-1v** was released with the paper [Language models enable zero-shot prediction of the effects of mutations on protein function](https://doi.org/10.1101/2021.07.09.450648) by Joshua Meier, Roshan Rao, Robert Verkuil, Jason Liu, Tom Sercu and Alexander Rives. **ESM-2 and ESMFold** were released with the paper [Language models of protein sequences at the scale of evolution enable accurate structure prediction](https://doi.org/10.1101/2022.07.20.500902) by Zeming Lin, Halil Akin, Roshan Rao, Brian Hie, Zhongkai Zhu, Wenting Lu, Allan dos Santos Costa, Maryam Fazel-Zarandi, Tom Sercu, Sal Candido, Alexander Rives.
1. **[FLAN-T5](model_doc/flan-t5)** (from Google AI) released in the repository [google-research/t5x](https://github.com/google-research/t5x/blob/main/docs/models.md#flan-t5-checkpoints) by Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Eric Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, Albert Webson, Shixiang Shane Gu, Zhuyun Dai, Mirac Suzgun, Xinyun Chen, Aakanksha Chowdhery, Sharan Narang, Gaurav Mishra, Adams Yu, Vincent Zhao, Yanping Huang, Andrew Dai, Hongkun Yu, Slav Petrov, Ed H. Chi, Jeff Dean, Jacob Devlin, Adam Roberts, Denny Zhou, Quoc V. Le, and Jason Wei
1. **[FlauBERT](model_doc/flaubert)** (from CNRS) released with the paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) by Hang Le, Loรฏc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoรฎt Crabbรฉ, Laurent Besacier, Didier Schwab.
1. **[FLAVA](model_doc/flava)** (from Facebook AI) released with the paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) by Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, and Douwe Kiela.
1. **[FNet](model_doc/fnet)** (from Google Research) released with the paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) by James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
1. **[Funnel Transformer](model_doc/funnel)** (from CMU/Google Brain) released with the paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) by Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
1. **[GIT](model_doc/git)** (from Microsoft Research) released with the paper [GIT: A Generative Image-to-text Transformer for Vision and Language](https://arxiv.org/abs/2205.14100) by Jianfeng Wang, Zhengyuan Yang, Xiaowei Hu, Linjie Li, Kevin Lin, Zhe Gan, Zicheng Liu, Ce Liu, Lijuan Wang.
1. **[GLPN](model_doc/glpn)** (from KAIST) released with the paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) by Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
1. **[GPT](model_doc/openai-gpt)** (from OpenAI) released with the paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) by Alec Radford, Karthik Narasimhan, Tim Salimans and Ilya Sutskever.
1. **[GPT Neo](model_doc/gpt_neo)** (from EleutherAI) released in the repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) by Sid Black, Stella Biderman, Leo Gao, Phil Wang and Connor Leahy.
1. **[GPT NeoX](model_doc/gpt_neox)** (from EleutherAI) released with the paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) by Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach
1. **[GPT NeoX Japanese](model_doc/gpt_neox_japanese)** (from ABEJA) released by Shinya Otani, Takayoshi Makabe, Anuj Arora, and Kyo Hattori.
1. **[GPT-2](model_doc/gpt2)** (from OpenAI) released with the paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) by Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** and Ilya Sutskever**.
1. **[GPT-J](model_doc/gptj)** (from EleutherAI) released in the repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) by Ben Wang and Aran Komatsuzaki.
1. **[GPT-Sw3](model_doc/gpt-sw3)** (from AI-Sweden) released with the paper [Lessons Learned from GPT-SW3: Building the First Large-Scale Generative Language Model for Swedish](http://www.lrec-conf.org/proceedings/lrec2022/pdf/2022.lrec-1.376.pdf) by Ariel Ekgren, Amaru Cuba Gyllensten, Evangelia Gogoulou, Alice Heiman, Severine Verlinden, Joey รhman, Fredrik Carlsson, Magnus Sahlgren.
1. **[Graphormer](model_doc/graphormer)** (from Microsoft) released with the paper [Do Transformers Really Perform Bad for Graph Representation?](https://arxiv.org/abs/2106.05234) by Chengxuan Ying, Tianle Cai, Shengjie Luo, Shuxin Zheng, Guolin Ke, Di He, Yanming Shen, Tie-Yan Liu.
1. **[GroupViT](model_doc/groupvit)** (from UCSD, NVIDIA) released with the paper [GroupViT: Semantic Segmentation Emerges from Text Supervision](https://arxiv.org/abs/2202.11094) by Jiarui Xu, Shalini De Mello, Sifei Liu, Wonmin Byeon, Thomas Breuel, Jan Kautz, Xiaolong Wang.
1. **[Hubert](model_doc/hubert)** (from Facebook) released with the paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) by Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
1. **[I-BERT](model_doc/ibert)** (from Berkeley) released with the paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) by Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
1. **[ImageGPT](model_doc/imagegpt)** (from OpenAI) released with the paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) by Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
1. **[Jukebox](model_doc/jukebox)** (from OpenAI) released with the paper [Jukebox: A Generative Model for Music](https://arxiv.org/pdf/2005.00341.pdf) by Prafulla Dhariwal, Heewoo Jun, Christine Payne, Jong Wook Kim, Alec Radford, Ilya Sutskever.
1. **[LayoutLM](model_doc/layoutlm)** (from Microsoft Research Asia) released with the paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) by Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
1. **[LayoutLMv2](model_doc/layoutlmv2)** (from Microsoft Research Asia) released with the paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) by Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
1. **[LayoutLMv3](model_doc/layoutlmv3)** (from Microsoft Research Asia) released with the paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) by Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
1. **[LayoutXLM](model_doc/layoutxlm)** (from Microsoft Research Asia) released with the paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) by Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
1. **[LED](model_doc/led)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
1. **[LeViT](model_doc/levit)** (from Meta AI) released with the paper [LeViT: A Vision Transformer in ConvNet's Clothing for Faster Inference](https://arxiv.org/abs/2104.01136) by Ben Graham, Alaaeldin El-Nouby, Hugo Touvron, Pierre Stock, Armand Joulin, Hervรฉ Jรฉgou, Matthijs Douze.
1. **[LiLT](model_doc/lilt)** (from South China University of Technology) released with the paper [LiLT: A Simple yet Effective Language-Independent Layout Transformer for Structured Document Understanding](https://arxiv.org/abs/2202.13669) by Jiapeng Wang, Lianwen Jin, Kai Ding.
1. **[Longformer](model_doc/longformer)** (from AllenAI) released with the paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) by Iz Beltagy, Matthew E. Peters, Arman Cohan.
1. **[LongT5](model_doc/longt5)** (from Google AI) released with the paper [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://arxiv.org/abs/2112.07916) by Mandy Guo, Joshua Ainslie, David Uthus, Santiago Ontanon, Jianmo Ni, Yun-Hsuan Sung, Yinfei Yang.
1. **[LUKE](model_doc/luke)** (from Studio Ousia) released with the paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) by Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
1. **[LXMERT](model_doc/lxmert)** (from UNC Chapel Hill) released with the paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) by Hao Tan and Mohit Bansal.
1. **[M-CTC-T](model_doc/mctct)** (from Facebook) released with the paper [Pseudo-Labeling For Massively Multilingual Speech Recognition](https://arxiv.org/abs/2111.00161) by Loren Lugosch, Tatiana Likhomanenko, Gabriel Synnaeve, and Ronan Collobert.
1. **[M2M100](model_doc/m2m_100)** (from Facebook) released with the paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) by Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
1. **[MarianMT](model_doc/marian)** Machine translation models trained using [OPUS](http://opus.nlpl.eu/) data by Jรถrg Tiedemann. The [Marian Framework](https://marian-nmt.github.io/) is being developed by the Microsoft Translator Team.
1. **[MarkupLM](model_doc/markuplm)** (from Microsoft Research Asia) released with the paper [MarkupLM: Pre-training of Text and Markup Language for Visually-rich Document Understanding](https://arxiv.org/abs/2110.08518) by Junlong Li, Yiheng Xu, Lei Cui, Furu Wei.
1. **[Mask2Former](model_doc/mask2former)** (from FAIR and UIUC) released with the paper [Masked-attention Mask Transformer for Universal Image Segmentation](https://arxiv.org/abs/2112.01527) by Bowen Cheng, Ishan Misra, Alexander G. Schwing, Alexander Kirillov, Rohit Girdhar.
1. **[MaskFormer](model_doc/maskformer)** (from Meta and UIUC) released with the paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) by Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
1. **[mBART](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) by Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
1. **[mBART-50](model_doc/mbart)** (from Facebook) released with the paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) by Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
1. **[Megatron-BERT](model_doc/megatron-bert)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
1. **[Megatron-GPT2](model_doc/megatron_gpt2)** (from NVIDIA) released with the paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) by Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper and Bryan Catanzaro.
1. **[mLUKE](model_doc/mluke)** (from Studio Ousia) released with the paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) by Ryokan Ri, Ikuya Yamada, and Yoshimasa Tsuruoka.
1. **[MobileBERT](model_doc/mobilebert)** (from CMU/Google Brain) released with the paper [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://arxiv.org/abs/2004.02984) by Zhiqing Sun, Hongkun Yu, Xiaodan Song, Renjie Liu, Yiming Yang, and Denny Zhou.
1. **[MobileNetV1](model_doc/mobilenet_v1)** (from Google Inc.) released with the paper [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://arxiv.org/abs/1704.04861) by Andrew G. Howard, Menglong Zhu, Bo Chen, Dmitry Kalenichenko, Weijun Wang, Tobias Weyand, Marco Andreetto, Hartwig Adam.
1. **[MobileNetV2](model_doc/mobilenet_v2)** (from Google Inc.) released with the paper [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://arxiv.org/abs/1801.04381) by Mark Sandler, Andrew Howard, Menglong Zhu, Andrey Zhmoginov, Liang-Chieh Chen.
1. **[MobileViT](model_doc/mobilevit)** (from Apple) released with the paper [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://arxiv.org/abs/2110.02178) by Sachin Mehta and Mohammad Rastegari.
1. **[MPNet](model_doc/mpnet)** (from Microsoft Research) released with the paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) by Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
1. **[MT5](model_doc/mt5)** (from Google AI) released with the paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) by Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
1. **[MVP](model_doc/mvp)** (from RUC AI Box) released with the paper [MVP: Multi-task Supervised Pre-training for Natural Language Generation](https://arxiv.org/abs/2206.12131) by Tianyi Tang, Junyi Li, Wayne Xin Zhao and Ji-Rong Wen.
1. **[NAT](model_doc/nat)** (from SHI Labs) released with the paper [Neighborhood Attention Transformer](https://arxiv.org/abs/2204.07143) by Ali Hassani, Steven Walton, Jiachen Li, Shen Li, and Humphrey Shi.
1. **[Nezha](model_doc/nezha)** (from Huawei Noahโs Ark Lab) released with the paper [NEZHA: Neural Contextualized Representation for Chinese Language Understanding](https://arxiv.org/abs/1909.00204) by Junqiu Wei, Xiaozhe Ren, Xiaoguang Li, Wenyong Huang, Yi Liao, Yasheng Wang, Jiashu Lin, Xin Jiang, Xiao Chen and Qun Liu.
1. **[NLLB](model_doc/nllb)** (from Meta) released with the paper [No Language Left Behind: Scaling Human-Centered Machine Translation](https://arxiv.org/abs/2207.04672) by the NLLB team.
1. **[Nystrรถmformer](model_doc/nystromformer)** (from the University of Wisconsin - Madison) released with the paper [Nystrรถmformer: A Nystrรถm-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) by Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
1. **[OneFormer](model_doc/oneformer)** (from SHI Labs) released with the paper [OneFormer: One Transformer to Rule Universal Image Segmentation](https://arxiv.org/abs/2211.06220) by Jitesh Jain, Jiachen Li, MangTik Chiu, Ali Hassani, Nikita Orlov, Humphrey Shi.
1. **[OPT](master/model_doc/opt)** (from Meta AI) released with the paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) by Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
1. **[OWL-ViT](model_doc/owlvit)** (from Google AI) released with the paper [Simple Open-Vocabulary Object Detection with Vision Transformers](https://arxiv.org/abs/2205.06230) by Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby.
1. **[Pegasus](model_doc/pegasus)** (from Google) released with the paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) by Jingqing Zhang, Yao Zhao, Mohammad Saleh and Peter J. Liu.
1. **[PEGASUS-X](model_doc/pegasus_x)** (from Google) released with the paper [Investigating Efficiently Extending Transformers for Long Input Summarization](https://arxiv.org/abs/2208.04347) by Jason Phang, Yao Zhao, and Peter J. Liu.
1. **[Perceiver IO](model_doc/perceiver)** (from Deepmind) released with the paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) by Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hรฉnaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, Joรฃo Carreira.
1. **[PhoBERT](model_doc/phobert)** (from VinAI Research) released with the paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) by Dat Quoc Nguyen and Anh Tuan Nguyen.
1. **[PLBart](model_doc/plbart)** (from UCLA NLP) released with the paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) by Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
1. **[PoolFormer](model_doc/poolformer)** (from Sea AI Labs) released with the paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) by Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng.
1. **[ProphetNet](model_doc/prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
1. **[QDQBert](model_doc/qdqbert)** (from NVIDIA) released with the paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) by Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev and Paulius Micikevicius.
1. **[RAG](model_doc/rag)** (from Facebook) released with the paper [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://arxiv.org/abs/2005.11401) by Patrick Lewis, Ethan Perez, Aleksandara Piktus, Fabio Petroni, Vladimir Karpukhin, Naman Goyal, Heinrich Kรผttler, Mike Lewis, Wen-tau Yih, Tim Rocktรคschel, Sebastian Riedel, Douwe Kiela.
1. **[REALM](model_doc/realm.html)** (from Google Research) released with the paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) by Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat and Ming-Wei Chang.
1. **[Reformer](model_doc/reformer)** (from Google Research) released with the paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) by Nikita Kitaev, ลukasz Kaiser, Anselm Levskaya.
1. **[RegNet](model_doc/regnet)** (from META Platforms) released with the paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) by Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollรกr.
1. **[RemBERT](model_doc/rembert)** (from Google Research) released with the paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) by Hyung Won Chung, Thibault Fรฉvry, Henry Tsai, M. Johnson, Sebastian Ruder.
1. **[ResNet](model_doc/resnet)** (from Microsoft Research) released with the paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) by Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.
1. **[RoBERTa](model_doc/roberta)** (from Facebook), released together with the paper [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) by Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
1. **[RoBERTa-PreLayerNorm](model_doc/roberta-prelayernorm)** (from Facebook) released with the paper [fairseq: A Fast, Extensible Toolkit for Sequence Modeling](https://arxiv.org/abs/1904.01038) by Myle Ott, Sergey Edunov, Alexei Baevski, Angela Fan, Sam Gross, Nathan Ng, David Grangier, Michael Auli.
1. **[RoCBert](model_doc/roc_bert)** (from WeChatAI) released with the paper [RoCBert: Robust Chinese Bert with Multimodal Contrastive Pretraining](https://aclanthology.org/2022.acl-long.65.pdf) by HuiSu, WeiweiShi, XiaoyuShen, XiaoZhou, TuoJi, JiaruiFang, JieZhou.
1. **[RoFormer](model_doc/roformer)** (from ZhuiyiTechnology), released together with the paper [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864) by Jianlin Su and Yu Lu and Shengfeng Pan and Bo Wen and Yunfeng Liu.
1. **[SegFormer](model_doc/segformer)** (from NVIDIA) released with the paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) by Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
1. **[SEW](model_doc/sew)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
1. **[SEW-D](model_doc/sew_d)** (from ASAPP) released with the paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) by Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
1. **[SpeechT5](model_doc/speecht5)** (from Microsoft Research) released with the paper [SpeechT5: Unified-Modal Encoder-Decoder Pre-Training for Spoken Language Processing](https://arxiv.org/abs/2110.07205) by Junyi Ao, Rui Wang, Long Zhou, Chengyi Wang, Shuo Ren, Yu Wu, Shujie Liu, Tom Ko, Qing Li, Yu Zhang, Zhihua Wei, Yao Qian, Jinyu Li, Furu Wei.
1. **[SpeechToTextTransformer](model_doc/speech_to_text)** (from Facebook), released together with the paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
1. **[SpeechToTextTransformer2](model_doc/speech_to_text_2)** (from Facebook), released together with the paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) by Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
1. **[Splinter](model_doc/splinter)** (from Tel Aviv University), released together with the paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) by Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
1. **[SqueezeBERT](model_doc/squeezebert)** (from Berkeley) released with the paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) by Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, and Kurt W. Keutzer.
1. **[Swin Transformer](model_doc/swin)** (from Microsoft) released with the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
1. **[Swin Transformer V2](model_doc/swinv2)** (from Microsoft) released with the paper [Swin Transformer V2: Scaling Up Capacity and Resolution](https://arxiv.org/abs/2111.09883) by Ze Liu, Han Hu, Yutong Lin, Zhuliang Yao, Zhenda Xie, Yixuan Wei, Jia Ning, Yue Cao, Zheng Zhang, Li Dong, Furu Wei, Baining Guo.
1. **[Swin2SR](model_doc/swin2sr)** (from University of Wรผrzburg) released with the paper [Swin2SR: SwinV2 Transformer for Compressed Image Super-Resolution and Restoration](https://arxiv.org/abs/2209.11345) by Marcos V. Conde, Ui-Jin Choi, Maxime Burchi, Radu Timofte.
1. **[SwitchTransformers](model_doc/switch_transformers)** (from Google) released with the paper [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://arxiv.org/abs/2101.03961) by William Fedus, Barret Zoph, Noam Shazeer.
1. **[T5](model_doc/t5)** (from Google AI) released with the paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
1. **[T5v1.1](model_doc/t5v1.1)** (from Google AI) released in the repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) by Colin Raffel and Noam Shazeer and Adam Roberts and Katherine Lee and Sharan Narang and Michael Matena and Yanqi Zhou and Wei Li and Peter J. Liu.
1. **[Table Transformer](model_doc/table-transformer)** (from Microsoft Research) released with the paper [PubTables-1M: Towards Comprehensive Table Extraction From Unstructured Documents](https://arxiv.org/abs/2110.00061) by Brandon Smock, Rohith Pesala, Robin Abraham.
1. **[TAPAS](model_doc/tapas)** (from Google AI) released with the paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) by Jonathan Herzig, Paweล Krzysztof Nowak, Thomas Mรผller, Francesco Piccinno and Julian Martin Eisenschlos.
1. **[TAPEX](model_doc/tapex)** (from Microsoft Research) released with the paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) by Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
1. **[Time Series Transformer](model_doc/time_series_transformer)** (from HuggingFace).
1. **[TimeSformer](model_doc/timesformer)** (from Facebook) released with the paper [Is Space-Time Attention All You Need for Video Understanding?](https://arxiv.org/abs/2102.05095) by Gedas Bertasius, Heng Wang, Lorenzo Torresani.
1. **[Trajectory Transformer](model_doc/trajectory_transformers)** (from the University of California at Berkeley) released with the paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) by Michael Janner, Qiyang Li, Sergey Levine
1. **[Transformer-XL](model_doc/transfo-xl)** (from Google/CMU) released with the paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) by Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
1. **[TrOCR](model_doc/trocr)** (from Microsoft), released together with the paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) by Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
1. **[UL2](model_doc/ul2)** (from Google Research) released with the paper [Unifying Language Learning Paradigms](https://arxiv.org/abs/2205.05131v1) by Yi Tay, Mostafa Dehghani, Vinh Q. Tran, Xavier Garcia, Dara Bahri, Tal Schuster, Huaixiu Steven Zheng, Neil Houlsby, Donald Metzler
1. **[UniSpeech](model_doc/unispeech)** (from Microsoft Research) released with the paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) by Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
1. **[UniSpeechSat](model_doc/unispeech-sat)** (from Microsoft Research) released with the paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) by Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
1. **[UPerNet](model_doc/upernet)** (from Peking University) released with the paper [Unified Perceptual Parsing for Scene Understanding](https://arxiv.org/abs/1807.10221) by Tete Xiao, Yingcheng Liu, Bolei Zhou, Yuning Jiang, Jian Sun.
1. **[VAN](model_doc/van)** (from Tsinghua University and Nankai University) released with the paper [Visual Attention Network](https://arxiv.org/abs/2202.09741) by Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
1. **[VideoMAE](model_doc/videomae)** (from Multimedia Computing Group, Nanjing University) released with the paper [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://arxiv.org/abs/2203.12602) by Zhan Tong, Yibing Song, Jue Wang, Limin Wang.
1. **[ViLT](model_doc/vilt)** (from NAVER AI Lab/Kakao Enterprise/Kakao Brain) released with the paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) by Wonjae Kim, Bokyung Son, Ildoo Kim.
1. **[Vision Transformer (ViT)](model_doc/vit)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
1. **[VisualBERT](model_doc/visual_bert)** (from UCLA NLP) released with the paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) by Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
1. **[ViT Hybrid](model_doc/vit_hybrid)** (from Google AI) released with the paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) by Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
1. **[ViTMAE](model_doc/vit_mae)** (from Meta AI) released with the paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) by Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollรกr, Ross Girshick.
1. **[ViTMSN](model_doc/vit_msn)** (from Meta AI) released with the paper [Masked Siamese Networks for Label-Efficient Learning](https://arxiv.org/abs/2204.07141) by Mahmoud Assran, Mathilde Caron, Ishan Misra, Piotr Bojanowski, Florian Bordes, Pascal Vincent, Armand Joulin, Michael Rabbat, Nicolas Ballas.
1. **[Wav2Vec2](model_doc/wav2vec2)** (from Facebook AI) released with the paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) by Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
1. **[Wav2Vec2-Conformer](model_doc/wav2vec2-conformer)** (from Facebook AI) released with the paper [FAIRSEQ S2T: Fast Speech-to-Text Modeling with FAIRSEQ](https://arxiv.org/abs/2010.05171) by Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Sravya Popuri, Dmytro Okhonko, Juan Pino.
1. **[Wav2Vec2Phoneme](model_doc/wav2vec2_phoneme)** (from Facebook AI) released with the paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) by Qiantong Xu, Alexei Baevski, Michael Auli.
1. **[WavLM](model_doc/wavlm)** (from Microsoft Research) released with the paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) by Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
1. **[Whisper](model_doc/whisper)** (from OpenAI) released with the paper [Robust Speech Recognition via Large-Scale Weak Supervision](https://cdn.openai.com/papers/whisper.pdf) by Alec Radford, Jong Wook Kim, Tao Xu, Greg Brockman, Christine McLeavey, Ilya Sutskever.
1. **[X-CLIP](model_doc/xclip)** (from Microsoft Research) released with the paper [Expanding Language-Image Pretrained Models for General Video Recognition](https://arxiv.org/abs/2208.02816) by Bolin Ni, Houwen Peng, Minghao Chen, Songyang Zhang, Gaofeng Meng, Jianlong Fu, Shiming Xiang, Haibin Ling.
1. **[XGLM](model_doc/xglm)** (From Facebook AI) released with the paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) by Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li.
1. **[XLM](model_doc/xlm)** (from Facebook) released together with the paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) by Guillaume Lample and Alexis Conneau.
1. **[XLM-ProphetNet](model_doc/xlm-prophetnet)** (from Microsoft Research) released with the paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) by Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang and Ming Zhou.
1. **[XLM-RoBERTa](model_doc/xlm-roberta)** (from Facebook AI), released together with the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmรกn, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov.
1. **[XLM-RoBERTa-XL](model_doc/xlm-roberta-xl)** (from Facebook AI), released together with the paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) by Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
1. **[XLNet](model_doc/xlnet)** (from Google/CMU) released with the paper [โXLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) by Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
1. **[XLS-R](model_doc/xls_r)** (from Facebook AI) released with the paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) by Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
1. **[XLSR-Wav2Vec2](model_doc/xlsr_wav2vec2)** (from Facebook AI) released with the paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) by Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
1. **[YOLOS](model_doc/yolos)** (from Huazhong University of Science & Technology) released with the paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) by Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
1. **[YOSO](model_doc/yoso)** (from the University of Wisconsin - Madison) released with the paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) by Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
### Frameworks compatibles
Le tableau ci-dessous reprรฉsente la prise en charge actuelle dans la bibliothรจque pour chacun de ces modรจles, qu'ils aient ou non un tokenizer Python (appelรฉ "slow"). Un tokenizer rapide ("fast") soutenu par la bibliothรจque ๐ค Tokenizers, qu'ils aient un support en Jax (via Flax), PyTorch, et/ou TensorFlow.
<!--This table is updated automatically from the auto modules with _make fix-copies_. Do not update manually!-->
| Modรจle | Tokenizer slow | Tokenizer fast | PyTorch support | TensorFlow support | Flax Support |
|:-----------------------------:|:--------------:|:--------------:|:---------------:|:------------------:|:------------:|
| ALBERT | โ
| โ
| โ
| โ
| โ
|
| AltCLIP | โ | โ | โ
| โ | โ |
| Audio Spectrogram Transformer | โ | โ | โ
| โ | โ |
| BART | โ
| โ
| โ
| โ
| โ
|
| BEiT | โ | โ | โ
| โ | โ
|
| BERT | โ
| โ
| โ
| โ
| โ
|
| Bert Generation | โ
| โ | โ
| โ | โ |
| BigBird | โ
| โ
| โ
| โ | โ
|
| BigBird-Pegasus | โ | โ | โ
| โ | โ |
| BioGpt | โ
| โ | โ
| โ | โ |
| BiT | โ | โ | โ
| โ | โ |
| Blenderbot | โ
| โ
| โ
| โ
| โ
|
| BlenderbotSmall | โ
| โ
| โ
| โ
| โ
|
| BLIP | โ | โ | โ
| โ | โ |
| BLOOM | โ | โ
| โ
| โ | โ |
| BridgeTower | โ | โ | โ
| โ | โ |
| CamemBERT | โ
| โ
| โ
| โ
| โ |
| CANINE | โ
| โ | โ
| โ | โ |
| Chinese-CLIP | โ | โ | โ
| โ | โ |
| CLIP | โ
| โ
| โ
| โ
| โ
|
| CLIPSeg | โ | โ | โ
| โ | โ |
| CodeGen | โ
| โ
| โ
| โ | โ |
| Conditional DETR | โ | โ | โ
| โ | โ |
| ConvBERT | โ
| โ
| โ
| โ
| โ |
| ConvNeXT | โ | โ | โ
| โ
| โ |
| CTRL | โ
| โ | โ
| โ
| โ |
| CvT | โ | โ | โ
| โ
| โ |
| Data2VecAudio | โ | โ | โ
| โ | โ |
| Data2VecText | โ | โ | โ
| โ | โ |
| Data2VecVision | โ | โ | โ
| โ
| โ |
| DeBERTa | โ
| โ
| โ
| โ
| โ |
| DeBERTa-v2 | โ
| โ
| โ
| โ
| โ |
| Decision Transformer | โ | โ | โ
| โ | โ |
| Deformable DETR | โ | โ | โ
| โ | โ |
| DeiT | โ | โ | โ
| โ
| โ |
| DETA | โ | โ | โ
| โ | โ |
| DETR | โ | โ | โ
| โ | โ |
| DiNAT | โ | โ | โ
| โ | โ |
| DistilBERT | โ
| โ
| โ
| โ
| โ
|
| DonutSwin | โ | โ | โ
| โ | โ |
| DPR | โ
| โ
| โ
| โ
| โ |
| DPT | โ | โ | โ
| โ | โ |
| EfficientFormer | โ | โ | โ
| โ | โ |
| ELECTRA | โ
| โ
| โ
| โ
| โ
|
| Encoder decoder | โ | โ | โ
| โ
| โ
|
| ERNIE | โ | โ | โ
| โ | โ |
| ESM | โ
| โ | โ
| โ
| โ |
| FairSeq Machine-Translation | โ
| โ | โ
| โ | โ |
| FlauBERT | โ
| โ | โ
| โ
| โ |
| FLAVA | โ | โ | โ
| โ | โ |
| FNet | โ
| โ
| โ
| โ | โ |
| Funnel Transformer | โ
| โ
| โ
| โ
| โ |
| GIT | โ | โ | โ
| โ | โ |
| GLPN | โ | โ | โ
| โ | โ |
| GPT Neo | โ | โ | โ
| โ | โ
|
| GPT NeoX | โ | โ
| โ
| โ | โ |
| GPT NeoX Japanese | โ
| โ | โ
| โ | โ |
| GPT-J | โ | โ | โ
| โ
| โ
|
| GPT-Sw3 | โ
| โ
| โ
| โ
| โ
|
| Graphormer | โ | โ | โ
| โ | โ |
| GroupViT | โ | โ | โ
| โ
| โ |
| Hubert | โ | โ | โ
| โ
| โ |
| I-BERT | โ | โ | โ
| โ | โ |
| ImageGPT | โ | โ | โ
| โ | โ |
| Jukebox | โ
| โ | โ
| โ | โ |
| LayoutLM | โ
| โ
| โ
| โ
| โ |
| LayoutLMv2 | โ
| โ
| โ
| โ | โ |
| LayoutLMv3 | โ
| โ
| โ
| โ
| โ |
| LED | โ
| โ
| โ
| โ
| โ |
| LeViT | โ | โ | โ
| โ | โ |
| LiLT | โ | โ | โ
| โ | โ |
| Longformer | โ
| โ
| โ
| โ
| โ |
| LongT5 | โ | โ | โ
| โ | โ
|
| LUKE | โ
| โ | โ
| โ | โ |
| LXMERT | โ
| โ
| โ
| โ
| โ |
| M-CTC-T | โ | โ | โ
| โ | โ |
| M2M100 | โ
| โ | โ
| โ | โ |
| Marian | โ
| โ | โ
| โ
| โ
|
| MarkupLM | โ
| โ
| โ
| โ | โ |
| Mask2Former | โ | โ | โ
| โ | โ |
| MaskFormer | โ | โ | โ
| โ | โ |
| MaskFormerSwin | โ | โ | โ | โ | โ |
| mBART | โ
| โ
| โ
| โ
| โ
|
| Megatron-BERT | โ | โ | โ
| โ | โ |
| MobileBERT | โ
| โ
| โ
| โ
| โ |
| MobileNetV1 | โ | โ | โ
| โ | โ |
| MobileNetV2 | โ | โ | โ
| โ | โ |
| MobileViT | โ | โ | โ
| โ
| โ |
| MPNet | โ
| โ
| โ
| โ
| โ |
| MT5 | โ
| โ
| โ
| โ
| โ
|
| MVP | โ
| โ
| โ
| โ | โ |
| NAT | โ | โ | โ
| โ | โ |
| Nezha | โ | โ | โ
| โ | โ |
| Nystrรถmformer | โ | โ | โ
| โ | โ |
| OneFormer | โ | โ | โ
| โ | โ |
| OpenAI GPT | โ
| โ
| โ
| โ
| โ |
| OpenAI GPT-2 | โ
| โ
| โ
| โ
| โ
|
| OPT | โ | โ | โ
| โ
| โ
|
| OWL-ViT | โ | โ | โ
| โ | โ |
| Pegasus | โ
| โ
| โ
| โ
| โ
|
| PEGASUS-X | โ | โ | โ
| โ | โ |
| Perceiver | โ
| โ | โ
| โ | โ |
| PLBart | โ
| โ | โ
| โ | โ |
| PoolFormer | โ | โ | โ
| โ | โ |
| ProphetNet | โ
| โ | โ
| โ | โ |
| QDQBert | โ | โ | โ
| โ | โ |
| RAG | โ
| โ | โ
| โ
| โ |
| REALM | โ
| โ
| โ
| โ | โ |
| Reformer | โ
| โ
| โ
| โ | โ |
| RegNet | โ | โ | โ
| โ
| โ
|
| RemBERT | โ
| โ
| โ
| โ
| โ |
| ResNet | โ | โ | โ
| โ
| โ |
| RetriBERT | โ
| โ
| โ
| โ | โ |
| RoBERTa | โ
| โ
| โ
| โ
| โ
|
| RoBERTa-PreLayerNorm | โ | โ | โ
| โ
| โ
|
| RoCBert | โ
| โ | โ
| โ | โ |
| RoFormer | โ
| โ
| โ
| โ
| โ
|
| SegFormer | โ | โ | โ
| โ
| โ |
| SEW | โ | โ | โ
| โ | โ |
| SEW-D | โ | โ | โ
| โ | โ |
| Speech Encoder decoder | โ | โ | โ
| โ | โ
|
| Speech2Text | โ
| โ | โ
| โ
| โ |
| Speech2Text2 | โ
| โ | โ | โ | โ |
| SpeechT5 | โ
| โ | โ
| โ | โ |
| Splinter | โ
| โ
| โ
| โ | โ |
| SqueezeBERT | โ
| โ
| โ
| โ | โ |
| Swin Transformer | โ | โ | โ
| โ
| โ |
| Swin Transformer V2 | โ | โ | โ
| โ | โ |
| Swin2SR | โ | โ | โ
| โ | โ |
| SwitchTransformers | โ | โ | โ
| โ | โ |
| T5 | โ
| โ
| โ
| โ
| โ
|
| Table Transformer | โ | โ | โ
| โ | โ |
| TAPAS | โ
| โ | โ
| โ
| โ |
| Time Series Transformer | โ | โ | โ
| โ | โ |
| TimeSformer | โ | โ | โ
| โ | โ |
| Trajectory Transformer | โ | โ | โ
| โ | โ |
| Transformer-XL | โ
| โ | โ
| โ
| โ |
| TrOCR | โ | โ | โ
| โ | โ |
| UniSpeech | โ | โ | โ
| โ | โ |
| UniSpeechSat | โ | โ | โ
| โ | โ |
| UPerNet | โ | โ | โ
| โ | โ |
| VAN | โ | โ | โ
| โ | โ |
| VideoMAE | โ | โ | โ
| โ | โ |
| ViLT | โ | โ | โ
| โ | โ |
| Vision Encoder decoder | โ | โ | โ
| โ
| โ
|
| VisionTextDualEncoder | โ | โ | โ
| โ | โ
|
| VisualBERT | โ | โ | โ
| โ | โ |
| ViT | โ | โ | โ
| โ
| โ
|
| ViT Hybrid | โ | โ | โ
| โ | โ |
| ViTMAE | โ | โ | โ
| โ
| โ |
| ViTMSN | โ | โ | โ
| โ | โ |
| Wav2Vec2 | โ
| โ | โ
| โ
| โ
|
| Wav2Vec2-Conformer | โ | โ | โ
| โ | โ |
| WavLM | โ | โ | โ
| โ | โ |
| Whisper | โ
| โ | โ
| โ
| โ |
| X-CLIP | โ | โ | โ
| โ | โ |
| XGLM | โ
| โ
| โ
| โ
| โ
|
| XLM | โ
| โ | โ
| โ
| โ |
| XLM-ProphetNet | โ
| โ | โ
| โ | โ |
| XLM-RoBERTa | โ
| โ
| โ
| โ
| โ
|
| XLM-RoBERTa-XL | โ | โ | โ
| โ | โ |
| XLNet | โ
| โ
| โ
| โ
| โ |
| YOLOS | โ | โ | โ
| โ | โ |
| YOSO | โ | โ | โ
| โ | โ |
<!-- End table-->
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/fr/quicktour.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Visite rapide
[[open-in-colab]]
Soyez opรฉrationnel avec ๐ค Transformers ! Que vous soyez un dรฉveloppeur ou un utilisateur lambda, cette visite rapide vous aidera ร dรฉmarrer et vous montrera comment utiliser le [`pipeline`] pour l'infรฉrence, charger un modรจle prรฉ-entraรฎnรฉ et un prรฉprocesseur avec une [AutoClass](./model_doc/auto), et entraรฎner rapidement un modรจle avec PyTorch ou TensorFlow. Si vous รชtes un dรฉbutant, nous vous recommandons de consulter nos tutoriels ou notre [cours](https://huggingface.co/course/chapter1/1) suivant pour des explications plus approfondies des concepts prรฉsentรฉs ici.
Avant de commencer, assurez-vous que vous avez installรฉ toutes les bibliothรจques nรฉcessaires :
```bash
!pip install transformers datasets
```
Vous aurez aussi besoin d'installer votre bibliothรจque d'apprentissage profond favorite :
<frameworkcontent>
<pt>
```bash
pip install torch
```
</pt>
<tf>
```bash
pip install tensorflow
```
</tf>
</frameworkcontent>
## Pipeline
<Youtube id="tiZFewofSLM"/>
Le [`pipeline`] est le moyen le plus simple d'utiliser un modรจle prรฉ-entraรฎnรฉ pour l'infรฉrence. Vous pouvez utiliser le [`pipeline`] prรชt ร l'emploi pour de nombreuses tรขches dans diffรฉrentes modalitรฉs. Consultez le tableau ci-dessous pour connaรฎtre les tรขches prises en charge :
| **Tรขche** | **Description** | **Modalitรฉ** | **Identifiant du pipeline** |
|------------------------------|--------------------------------------------------------------------------------------------------------------|----------------------|-----------------------------------------------|
| Classification de texte | Attribue une catรฉgorie ร une sรฉquence de texte donnรฉe | Texte | pipeline(task="sentiment-analysis") |
| Gรฉnรฉration de texte | Gรฉnรจre du texte ร partir d'une consigne donnรฉe | Texte | pipeline(task="text-generation") |
| Reconnaissance de token nommรฉ | Attribue une catรฉgorie ร chaque token dans une sรฉquence (personnes, organisation, localisation, etc.) | Texte | pipeline(task="ner") |
| Question rรฉponse | Extrait une rรฉponse du texte en fonction du contexte et d'une question | Texte | pipeline(task="question-answering") |
| Prรฉdiction de token masquรฉ | Prรฉdit correctement le token masquรฉ dans une sรฉquence | Texte | pipeline(task="fill-mask") |
| Gรฉnรฉration de rรฉsumรฉ | Gรฉnรจre un rรฉsumรฉ d'une sรฉquence de texte donnรฉe ou d'un document | Texte | pipeline(task="summarization") |
| Traduction | Traduit du texte d'un langage ร un autre | Texte | pipeline(task="translation") |
| Classification d'image | Attribue une catรฉgorie ร une image | Image | pipeline(task="image-classification") |
| Segmentation d'image | Attribue une catรฉgorie ร chaque pixel d'une image (supporte la segmentation sรฉmantique, panoptique et d'instance) | Image | pipeline(task="image-segmentation") |
| Dรฉtection d'objets | Prรฉdit les dรฉlimitations et catรฉgories d'objets dans une image | Image | pipeline(task="object-detection") |
| Classification d'audio | Attribue une catรฉgorie ร un fichier audio | Audio | pipeline(task="audio-classification") |
| Reconnaissance automatique de la parole | Extrait le discours d'un fichier audio en texte | Audio | pipeline(task="automatic-speech-recognition") |
| Question rรฉponse visuels | Etant donnรฉes une image et une question, rรฉpond correctement ร une question sur l'image | Modalitรฉs multiples | pipeline(task="vqa") |
Commencez par crรฉer une instance de [`pipeline`] et spรฉcifiez la tรขche pour laquelle vous souhaitez l'utiliser. Vous pouvez utiliser le [`pipeline`] pour n'importe laquelle des tรขches mentionnรฉes dans le tableau prรฉcรฉdent. Pour obtenir une liste complรจte des tรขches prises en charge, consultez la documentation de l'[API pipeline](./main_classes/pipelines). Dans ce guide, nous utiliserons le [`pipeline`] pour l'analyse des sentiments ร titre d'exemple :
```py
>>> from transformers import pipeline
>>> classifier = pipeline("sentiment-analysis")
```
Le [`pipeline`] tรฉlรฉcharge et stocke en cache un [modรจle prรฉ-entraรฎnรฉ](https://huggingface.co/distilbert-base-uncased-finetuned-sst-2-english) et un tokenizer par dรฉfaut pour l'analyse des sentiments. Vous pouvez maintenant utiliser le `classifier` sur le texte de votre choix :
```py
>>> classifier("We are very happy to show you the ๐ค Transformers library.")
[{'label': 'POSITIVE', 'score': 0.9998}]
```
Si vous voulez classifier plus qu'un texte, donnez une liste de textes au [`pipeline`] pour obtenir une liste de dictionnaires en retour :
```py
>>> results = classifier(["We are very happy to show you the ๐ค Transformers library.", "We hope you don't hate it."])
>>> for result in results:
... print(f"label: {result['label']}, avec le score de: {round(result['score'], 4)}")
label: POSITIVE, avec le score de: 0.9998
label: NEGATIVE, avec le score de: 0.5309
```
Le [`pipeline`] peut aussi itรฉrer sur un jeu de donnรฉes entier pour n'importe quelle tรขche. Prenons par exemple la reconnaissance automatique de la parole :
```py
>>> import torch
>>> from transformers import pipeline
>>> speech_recognizer = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-base-960h")
```
Chargez un jeu de donnรฉes audio (voir le ๐ค Datasets [Quick Start](https://huggingface.co/docs/datasets/quickstart#audio) pour plus de dรฉtails) sur lequel vous souhaitez itรฉrer. Pour cet exemple, nous chargeons le jeu de donnรฉes [MInDS-14](https://huggingface.co/datasets/PolyAI/minds14) :
```py
>>> from datasets import load_dataset, Audio
>>> dataset = load_dataset("PolyAI/minds14", name="en-US", split="train") # doctest: +IGNORE_RESULT
```
Vous devez vous assurer que le taux d'รฉchantillonnage de l'ensemble de donnรฉes correspond au taux d'รฉchantillonnage sur lequel [`facebook/wav2vec2-base-960h`](https://huggingface.co/facebook/wav2vec2-base-960h) a รฉtรฉ entraรฎnรฉ :
```py
>>> dataset = dataset.cast_column("audio", Audio(sampling_rate=speech_recognizer.feature_extractor.sampling_rate))
```
Les fichiers audio sont automatiquement chargรฉs et rรฉรฉchantillonnรฉs lors de l'appel de la colonne `"audio"`.
Extrayez les tableaux de formes d'ondes brutes des quatre premiers รฉchantillons et passez-les comme une liste au pipeline :
```py
>>> result = speech_recognizer(dataset[:4]["audio"])
>>> print([d["text"] for d in result])
['I WOULD LIKE TO SET UP A JOINT ACCOUNT WITH MY PARTNER HOW DO I PROCEED WITH DOING THAT', "FODING HOW I'D SET UP A JOIN TO HET WITH MY WIFE AND WHERE THE AP MIGHT BE", "I I'D LIKE TOY SET UP A JOINT ACCOUNT WITH MY PARTNER I'M NOT SEEING THE OPTION TO DO IT ON THE AP SO I CALLED IN TO GET SOME HELP CAN I JUST DO IT OVER THE PHONE WITH YOU AND GIVE YOU THE INFORMATION OR SHOULD I DO IT IN THE AP AND I'M MISSING SOMETHING UQUETTE HAD PREFERRED TO JUST DO IT OVER THE PHONE OF POSSIBLE THINGS", 'HOW DO I THURN A JOIN A COUNT']
```
Pour les ensembles de donnรฉes plus importants oรน les entrรฉes sont volumineuses (comme dans les domaines de la parole ou de la vision), utilisez plutรดt un gรฉnรฉrateur au lieu d'une liste pour charger toutes les entrรฉes en mรฉmoire. Pour plus d'informations, consultez la documentation de l'[API pipeline](./main_classes/pipelines).
### Utiliser une autre modรจle et tokenizer dans le pipeline
Le [`pipeline`] peut รชtre utilisรฉ avec n'importe quel modรจle du [Hub](https://huggingface.co/models), ce qui permet d'adapter facilement le [`pipeline`] ร d'autres cas d'utilisation. Par exemple, si vous souhaitez un modรจle capable de traiter du texte franรงais, utilisez les filtres du Hub pour trouver un modรจle appropriรฉ. Le premier rรฉsultat renvoie un [modรจle BERT](https://huggingface.co/nlptown/bert-base-multilingual-uncased-sentiment) multilingue finetunรฉ pour l'analyse des sentiments que vous pouvez utiliser pour le texte franรงais :
```py
>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
```
<frameworkcontent>
<pt>
Utilisez [`AutoModelForSequenceClassification`] et [`AutoTokenizer`] pour charger le modรจle prรฉ-entraรฎnรฉ et le tokenizer adaptรฉ (plus de dรฉtails sur une `AutoClass` dans la section suivante) :
```py
>>> from transformers import AutoTokenizer, AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained(model_name)
>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
```
</pt>
<tf>
Utilisez [`TFAutoModelForSequenceClassification`] et [`AutoTokenizer`] pour charger le modรจle prรฉ-entraรฎnรฉ et le tokenizer adaptรฉ (plus de dรฉtails sur une `TFAutoClass` dans la section suivante) :
```py
>>> from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
>>> model = TFAutoModelForSequenceClassification.from_pretrained(model_name)
>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
```
</tf>
</frameworkcontent>
Spรฉcifiez le modรจle et le tokenizer dans le [`pipeline`], et utilisez le `classifier` sur le texte en franรงais :
```py
>>> classifier = pipeline("sentiment-analysis", model=model, tokenizer=tokenizer)
>>> classifier("Nous sommes trรจs heureux de vous prรฉsenter la bibliothรจque ๐ค Transformers.")
[{'label': '5 stars', 'score': 0.7273}]
```
Si vous ne parvenez pas ร trouver un modรจle adaptรฉ ร votre cas d'utilisation, vous devrez finetuner un modรจle prรฉ-entraรฎnรฉ sur vos donnรฉes. Jetez un coup d'ลil ร notre [tutoriel sur le finetuning](./training) pour apprendre comment faire. Enfin, aprรจs avoir finetunรฉ votre modรจle prรฉ-entraรฎnรฉ, pensez ร [partager](./model_sharing) le modรจle avec la communautรฉ sur le Hub afin de dรฉmocratiser l'apprentissage automatique pour tous ! ๐ค
## AutoClass
<Youtube id="AhChOFRegn4"/>
Les classes [`AutoModelForSequenceClassification`] et [`AutoTokenizer`] fonctionnent ensemble pour crรฉer un [`pipeline`] comme celui que vous avez utilisรฉ ci-dessus. Une [AutoClass](./model_doc/auto) est un raccourci qui rรฉcupรจre automatiquement l'architecture d'un modรจle prรฉ-entraรฎnรฉ ร partir de son nom ou de son emplacement. Il vous suffit de sรฉlectionner l'`AutoClass` appropriรฉe ร votre tรขche et la classe de prรฉtraitement qui lui est associรฉe.
Reprenons l'exemple de la section prรฉcรฉdente et voyons comment vous pouvez utiliser l'`AutoClass` pour reproduire les rรฉsultats du [`pipeline`].
### AutoTokenizer
Un tokenizer est chargรฉ de prรฉtraiter le texte pour en faire un tableau de chiffres qui servira d'entrรฉe ร un modรจle. De nombreuses rรจgles rรฉgissent le processus de tokenisation, notamment la maniรจre de diviser un mot et le niveau auquel les mots doivent รชtre divisรฉs (pour en savoir plus sur la tokenisation, consultez le [rรฉsumรฉ](./tokenizer_summary)). La chose la plus importante ร retenir est que vous devez instancier un tokenizer avec le mรชme nom de modรจle pour vous assurer que vous utilisez les mรชmes rรจgles de tokenisation que celles avec lesquelles un modรจle a รฉtรฉ prรฉ-entraรฎnรฉ.
Chargez un tokenizer avec [`AutoTokenizer`] :
```py
>>> from transformers import AutoTokenizer
>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
```
Passez votre texte au tokenizer :
```py
>>> encoding = tokenizer("We are very happy to show you the ๐ค Transformers library.")
>>> print(encoding)
{'input_ids': [101, 11312, 10320, 12495, 19308, 10114, 11391, 10855, 10103, 100, 58263, 13299, 119, 102],
'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}
```
Le tokenizer retourne un dictionnaire contenant :
* [input_ids](./glossary#input-ids): la reprรฉsentation numรฉrique des tokens.
* [attention_mask](.glossary#attention-mask): indique quels tokens doivent faire l'objet d'une attention particuliรจre (plus particuliรจrement les tokens de remplissage).
Un tokenizer peut รฉgalement accepter une liste de textes, et remplir et tronquer le texte pour retourner un รฉchantillon de longueur uniforme :
<frameworkcontent>
<pt>
```py
>>> pt_batch = tokenizer(
... ["We are very happy to show you the ๐ค Transformers library.", "We hope you don't hate it."],
... padding=True,
... truncation=True,
... max_length=512,
... return_tensors="pt",
... )
```
</pt>
<tf>
```py
>>> tf_batch = tokenizer(
... ["We are very happy to show you the ๐ค Transformers library.", "We hope you don't hate it."],
... padding=True,
... truncation=True,
... max_length=512,
... return_tensors="tf",
... )
```
</tf>
</frameworkcontent>
<Tip>
Consultez le tutoriel [prรฉtraitement](./preprocessing) pour plus de dรฉtails sur la tokenisation, et sur la maniรจre d'utiliser un [`AutoImageProcessor`], un [`AutoFeatureExtractor`] et un [`AutoProcessor`] pour prรฉtraiter les images, l'audio et les contenus multimodaux.
</Tip>
### AutoModel
<frameworkcontent>
<pt>
๐ค Transformers fournit un moyen simple et unifiรฉ de charger des instances prรฉ-entraรฎnรฉes. Cela signifie que vous pouvez charger un [`AutoModel`] comme vous chargeriez un [`AutoTokenizer`]. La seule diffรฉrence est de sรฉlectionner l'[`AutoModel`] appropriรฉ pour la tรขche. Pour une classification de texte (ou de sรฉquence de textes), vous devez charger [`AutoModelForSequenceClassification`] :
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
>>> pt_model = AutoModelForSequenceClassification.from_pretrained(model_name)
```
<Tip>
Voir le [rรฉsumรฉ de la tรขche](./task_summary) pour vรฉrifier si elle est prise en charge par une classe [`AutoModel`].
</Tip>
Maintenant, passez votre รฉchantillon d'entrรฉes prรฉtraitรฉes directement au modรจle. Il vous suffit de dรฉcompresser le dictionnaire en ajoutant `**` :
```py
>>> pt_outputs = pt_model(**pt_batch)
```
Le modรจle produit les activations finales dans l'attribut `logits`. Appliquez la fonction softmax aux `logits` pour rรฉcupรฉrer les probabilitรฉs :
```py
>>> from torch import nn
>>> pt_predictions = nn.functional.softmax(pt_outputs.logits, dim=-1)
>>> print(pt_predictions)
tensor([[0.0021, 0.0018, 0.0115, 0.2121, 0.7725],
[0.2084, 0.1826, 0.1969, 0.1755, 0.2365]], grad_fn=<SoftmaxBackward0>)
```
</pt>
<tf>
๐ค Transformers fournit un moyen simple et unifiรฉ de charger des instances prรฉ-entraรฎnรฉs. Cela signifie que vous pouvez charger un [`TFAutoModel`] comme vous chargeriez un [`AutoTokenizer`]. La seule diffรฉrence est de sรฉlectionner le [`TFAutoModel`] appropriรฉ pour la tรขche. Pour une classification de texte (ou de sรฉquence de textes), vous devez charger [`TFAutoModelForSequenceClassification`] :
```py
>>> from transformers import TFAutoModelForSequenceClassification
>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained(model_name)
```
<Tip>
Voir le [rรฉsumรฉ de la tรขche](./task_summary) pour vรฉrifier si elle est prise en charge par une classe [`AutoModel`].
</Tip>
Passez maintenant votre รฉchantillon d'entrรฉes prรฉtraitรฉes directement au modรจle en passant les clรฉs du dictionnaire directement aux tensors :
```py
>>> tf_outputs = tf_model(tf_batch)
```
Le modรจle produit les activations finales dans l'attribut `logits`. Appliquez la fonction softmax aux `logits` pour rรฉcupรฉrer les probabilitรฉs :
```py
>>> import tensorflow as tf
>>> tf_predictions = tf.nn.softmax(tf_outputs.logits, axis=-1)
>>> tf_predictions # doctest: +IGNORE_RESULT
```
</tf>
</frameworkcontent>
<Tip>
Tous les modรจles ๐ค Transformers (PyTorch ou TensorFlow) produisent les tensors *avant* la fonction d'activation finale (comme softmax) car la fonction d'activation finale est souvent fusionnรฉe avec le calcul de la perte. Les structures produites par le modรจle sont des classes de donnรฉes spรฉciales, de sorte que leurs attributs sont autocomplรฉtรฉs dans un environnement de dรฉveloppement. Les structures produites par le modรจle se comportent comme un tuple ou un dictionnaire (vous pouvez les indexer avec un entier, une tranche ou une chaรฎne), auquel cas les attributs qui sont None sont ignorรฉs.
</Tip>
### Sauvegarder un modรจle
<frameworkcontent>
<pt>
Une fois que votre modรจle est finetunรฉ, vous pouvez le sauvegarder avec son tokenizer en utilisant [`PreTrainedModel.save_pretrained`] :
```py
>>> pt_save_directory = "./pt_save_pretrained"
>>> tokenizer.save_pretrained(pt_save_directory) # doctest: +IGNORE_RESULT
>>> pt_model.save_pretrained(pt_save_directory)
```
Lorsque vous voulez rรฉutiliser le modรจle, rechargez-le avec [`PreTrainedModel.from_pretrained`] :
```py
>>> pt_model = AutoModelForSequenceClassification.from_pretrained("./pt_save_pretrained")
```
</pt>
<tf>
Une fois que votre modรจle est finetunรฉ, vous pouvez le sauvegarder avec son tokenizer en utilisant [`TFPreTrainedModel.save_pretrained`] :
```py
>>> tf_save_directory = "./tf_save_pretrained"
>>> tokenizer.save_pretrained(tf_save_directory) # doctest: +IGNORE_RESULT
>>> tf_model.save_pretrained(tf_save_directory)
```
Lorsque vous voulez rรฉutiliser le modรจle, rechargez-le avec [`TFPreTrainedModel.from_pretrained`] :
```py
>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained("./tf_save_pretrained")
```
</tf>
</frameworkcontent>
Une fonctionnalitรฉ particuliรจrement cool ๐ค Transformers est la possibilitรฉ d'enregistrer un modรจle et de le recharger en tant que modรจle PyTorch ou TensorFlow. Le paramรจtre `from_pt` ou `from_tf` permet de convertir le modรจle d'un framework ร l'autre :
<frameworkcontent>
<pt>
```py
>>> from transformers import AutoModel
>>> tokenizer = AutoTokenizer.from_pretrained(tf_save_directory)
>>> pt_model = AutoModelForSequenceClassification.from_pretrained(tf_save_directory, from_tf=True)
```
</pt>
<tf>
```py
>>> from transformers import TFAutoModel
>>> tokenizer = AutoTokenizer.from_pretrained(pt_save_directory)
>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained(pt_save_directory, from_pt=True)
```
</tf>
</frameworkcontent>
## Constructions de modรจles personnalisรฉs
Vous pouvez modifier la configuration du modรจle pour changer la faรงon dont un modรจle est construit. La configuration spรฉcifie les attributs d'un modรจle, tels que le nombre de couches ou de tรชtes d'attention. Vous partez de zรฉro lorsque vous initialisez un modรจle ร partir d'une configuration personnalisรฉe. Les attributs du modรจle sont initialisรฉs de maniรจre alรฉatoire et vous devrez entraรฎner le modรจle avant de pouvoir l'utiliser pour obtenir des rรฉsultats significatifs.
Commencez par importer [`AutoConfig`], puis chargez le modรจle prรฉ-entraรฎnรฉ que vous voulez modifier. Dans [`AutoConfig.from_pretrained`], vous pouvez spรฉcifier l'attribut que vous souhaitez modifier, tel que le nombre de tรชtes d'attention :
```py
>>> from transformers import AutoConfig
>>> my_config = AutoConfig.from_pretrained("distilbert-base-uncased", n_heads=12)
```
<frameworkcontent>
<pt>
Crรฉez un modรจle personnalisรฉ ร partir de votre configuration avec [`AutoModel.from_config`] :
```py
>>> from transformers import AutoModel
>>> my_model = AutoModel.from_config(my_config)
```
</pt>
<tf>
Crรฉez un modรจle personnalisรฉ ร partir de votre configuration avec [`TFAutoModel.from_config`] :
```py
>>> from transformers import TFAutoModel
>>> my_model = TFAutoModel.from_config(my_config)
```
</tf>
</frameworkcontent>
Consultez le guide [Crรฉer une architecture personnalisรฉe](./create_a_model) pour plus d'informations sur la crรฉation de configurations personnalisรฉes.
## Trainer - une boucle d'entraรฎnement optimisรฉe par PyTorch
Tous les modรจles sont des [`torch.nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) standard, vous pouvez donc les utiliser dans n'importe quelle boucle d'entraรฎnement typique. Bien que vous puissiez รฉcrire votre propre boucle d'entraรฎnement, ๐ค Transformers fournit une classe [`Trainer`] pour PyTorch, qui contient la boucle d'entraรฎnement de base et ajoute des fonctionnalitรฉs supplรฉmentaires comme l'entraรฎnement distribuรฉ, la prรฉcision mixte, et plus encore.
En fonction de votre tรขche, vous passerez gรฉnรฉralement les paramรจtres suivants ร [`Trainer`] :
1. Un [`PreTrainedModel`] ou un [`torch.nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module):
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased")
```
2. [`TrainingArguments`] contient les hyperparamรจtres du modรจle que vous pouvez changer comme le taux d'apprentissage, la taille de l'รฉchantillon, et le nombre d'รฉpoques pour s'entraรฎner. Les valeurs par dรฉfaut sont utilisรฉes si vous ne spรฉcifiez pas d'hyperparamรจtres d'apprentissage :
```py
>>> from transformers import TrainingArguments
>>> training_args = TrainingArguments(
... output_dir="path/to/save/folder/",
... learning_rate=2e-5,
... per_device_train_batch_size=8,
... per_device_eval_batch_size=8,
... num_train_epochs=2,
... )
```
3. Une classe de prรฉtraitement comme un tokenizer, un processeur d'images ou un extracteur de caractรฉristiques :
```py
>>> from transformers import AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased")
```
4. Chargez un jeu de donnรฉes :
```py
>>> from datasets import load_dataset
>>> dataset = load_dataset("rotten_tomatoes") # doctest: +IGNORE_RESULT
```
5. Crรฉez une fonction qui transforme le texte du jeu de donnรฉes en token :
```py
>>> def tokenize_dataset(dataset):
... return tokenizer(dataset["text"])
```
Puis appliquez-la ร l'intรฉgralitรฉ du jeu de donnรฉes avec [`~datasets.Dataset.map`]:
```py
>>> dataset = dataset.map(tokenize_dataset, batched=True)
```
6. Un [`DataCollatorWithPadding`] pour crรฉer un รฉchantillon d'exemples ร partir de votre jeu de donnรฉes :
```py
>>> from transformers import DataCollatorWithPadding
>>> data_collator = DataCollatorWithPadding(tokenizer=tokenizer)
```
Maintenant, rassemblez tous ces รฉlรฉments dans un [`Trainer`] :
```py
>>> from transformers import Trainer
>>> trainer = Trainer(
... model=model,
... args=training_args,
... train_dataset=dataset["train"],
... eval_dataset=dataset["test"],
... tokenizer=tokenizer,
... data_collator=data_collator,
... ) # doctest: +SKIP
```
Une fois que vous รชtes prรชt, appelez la fonction [`~Trainer.train`] pour commencer l'entraรฎnement :
```py
>>> trainer.train() # doctest: +SKIP
```
<Tip>
Pour les tรขches - comme la traduction ou la gรฉnรฉration de rรฉsumรฉ - qui utilisent un modรจle sรฉquence ร sรฉquence, utilisez plutรดt les classes [`Seq2SeqTrainer`] et [`Seq2SeqTrainingArguments`].
</Tip>
Vous pouvez personnaliser le comportement de la boucle d'apprentissage en redรฉfinissant les mรฉthodes ร l'intรฉrieur de [`Trainer`]. Cela vous permet de personnaliser des caractรฉristiques telles que la fonction de perte, l'optimiseur et le planificateur. Consultez la documentation de [`Trainer`] pour savoir quelles mรฉthodes peuvent รชtre redรฉfinies.
L'autre moyen de personnaliser la boucle d'apprentissage est d'utiliser les [Callbacks](./main_classes/callbacks). Vous pouvez utiliser les callbacks pour intรฉgrer d'autres bibliothรจques et inspecter la boucle d'apprentissage afin de suivre la progression ou d'arrรชter l'apprentissage plus tรดt. Les callbacks ne modifient rien dans la boucle d'apprentissage elle-mรชme. Pour personnaliser quelque chose comme la fonction de perte, vous devez redรฉfinir le [`Trainer`] ร la place.
## Entraรฎnement avec TensorFlow
Tous les modรจles sont des modรจles standard [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) afin qu'ils puissent รชtre entraรฎnรฉs avec TensorFlow avec l'API [Keras](https://keras.io/). ๐ค Transformers fournit la fonction [`~TFPreTrainedModel.prepare_tf_dataset`] pour charger facilement votre jeu de donnรฉes comme un `tf.data.Dataset` afin que vous puissiez commencer l'entraรฎnement immรฉdiatement avec les fonctions [`compile`](https://keras.io/api/models/model_training_apis/#compile-method) et [`fit`](https://keras.io/api/models/model_training_apis/#fit-method) de Keras.
1. Vous commencez avec un modรจle [`TFPreTrainedModel`] ou [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) :
```py
>>> from transformers import TFAutoModelForSequenceClassification
>>> model = TFAutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased")
```
2. Une classe de prรฉtraitement comme un tokenizer, un processeur d'images ou un extracteur de caractรฉristiques :
```py
>>> from transformers import AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased")
```
3. Crรฉez une fonction qui transforme le texte du jeu de donnรฉes en token :
```py
>>> def tokenize_dataset(dataset):
... return tokenizer(dataset["text"]) # doctest: +SKIP
```
4. Appliquez le tokenizer ร l'ensemble du jeu de donnรฉes avec [`~datasets.Dataset.map`] et passez ensuite le jeu de donnรฉes et le tokenizer ร [`~TFPreTrainedModel.prepare_tf_dataset`]. Vous pouvez รฉgalement modifier la taille de l'รฉchantillon et mรฉlanger le jeu de donnรฉes ici si vous le souhaitez :
```py
>>> dataset = dataset.map(tokenize_dataset) # doctest: +SKIP
>>> tf_dataset = model.prepare_tf_dataset(
... dataset, batch_size=16, shuffle=True, tokenizer=tokenizer
... ) # doctest: +SKIP
```
5. Une fois que vous รชtes prรชt, appelez les fonctions `compile` et `fit` pour commencer l'entraรฎnement :
```py
>>> from tensorflow.keras.optimizers import Adam
>>> model.compile(optimizer=Adam(3e-5))
>>> model.fit(dataset) # doctest: +SKIP
```
## Et aprรจs ?
Maintenant que vous avez terminรฉ la visite rapide de ๐ค Transformers, consultez nos guides et apprenez ร faire des choses plus spรฉcifiques comme crรฉer un modรจle personnalisรฉ, finetuner un modรจle pour une tรขche, et comment entraรฎner un modรจle avec un script. Si vous souhaitez en savoir plus sur les concepts fondamentaux de ๐ค Transformers, jetez un ลil ร nos guides conceptuels !
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/multilingual.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Modelli multilingue per l'inferenza
[[open-in-colab]]
Ci sono diversi modelli multilingue in ๐ค Transformers, e il loro utilizzo per l'inferenza differisce da quello dei modelli monolingua. Non *tutti* gli utilizzi dei modelli multilingue sono perรฒ diversi. Alcuni modelli, come [bert-base-multilingual-uncased](https://huggingface.co/bert-base-multilingual-uncased), possono essere usati come un modello monolingua. Questa guida ti mostrerร come utilizzare modelli multilingue che utilizzano un modo diverso per fare l'inferenza.
## XLM
XLM ha dieci diversi checkpoint, di cui solo uno รจ monolingua. I nove checkpoint rimanenti possono essere suddivisi in due categorie: i checkpoint che utilizzano i language embeddings e quelli che non li utilizzano.
### XLM con language embeddings
I seguenti modelli XLM utilizzano gli embeddings linguistici per specificare la lingua utilizzata per l'inferenza:
- `xlm-mlm-ende-1024` (Modellazione mascherata del linguaggio (Masked language modeling, in inglese), Inglese-Tedesco)
- `xlm-mlm-enfr-1024` (Modellazione mascherata del linguaggio, Inglese-Francese)
- `xlm-mlm-enro-1024` (Modellazione mascherata del linguaggio, Inglese-Rumeno)
- `xlm-mlm-xnli15-1024` (Modellazione mascherata del linguaggio, lingue XNLI)
- `xlm-mlm-tlm-xnli15-1024` (Modellazione mascherata del linguaggio + traduzione, lingue XNLI)
- `xlm-clm-enfr-1024` (Modellazione causale del linguaggio, Inglese-Francese)
- `xlm-clm-ende-1024` (Modellazione causale del linguaggio, Inglese-Tedesco)
Gli embeddings linguistici sono rappresentati come un tensore delle stesse dimensioni dell' `input_ids` passato al modello. I valori in questi tensori dipendono dal linguaggio usato e sono identificati dagli attributi `lang2id` e `id2lang` del tokenizer.
In questo esempio, carica il checkpoint `xlm-clm-enfr-1024` (Modellazione causale del linguaggio, Inglese-Francese):
```py
>>> import torch
>>> from transformers import XLMTokenizer, XLMWithLMHeadModel
>>> tokenizer = XLMTokenizer.from_pretrained("xlm-clm-enfr-1024")
>>> model = XLMWithLMHeadModel.from_pretrained("xlm-clm-enfr-1024")
```
L'attributo `lang2id` del tokenizer mostra il linguaggio del modello e il suo ids:
```py
>>> print(tokenizer.lang2id)
{'en': 0, 'fr': 1}
```
Poi, crea un esempio di input:
```py
>>> input_ids = torch.tensor([tokenizer.encode("Wikipedia was used to")]) # batch size of 1
```
Imposta l'id del linguaggio a `"en"` e usalo per definire il language embedding. Il language embedding รจ un tensore riempito con `0` perchรฉ questo รจ il language id per l'inglese. Questo tensore dovrebbe avere la stessa dimensione di `input_ids`.
```py
>>> language_id = tokenizer.lang2id["en"] # 0
>>> langs = torch.tensor([language_id] * input_ids.shape[1]) # torch.tensor([0, 0, 0, ..., 0])
>>> # We reshape it to be of size (batch_size, sequence_length)
>>> langs = langs.view(1, -1) # is now of shape [1, sequence_length] (we have a batch size of 1)
```
Adesso puoi inserire `input_ids` e language embedding nel modello:
```py
>>> outputs = model(input_ids, langs=langs)
```
Lo script [run_generation.py](https://github.com/huggingface/transformers/tree/main/examples/pytorch/text-generation/run_generation.py) puรฒ generare testo tramite i language embeddings usando i checkpoints `xlm-clm`.
### XLM senza language embeddings
I seguenti modelli XLM non richiedono l'utilizzo dei language embeddings per fare inferenza:
- `xlm-mlm-17-1280` (Modellazione mascherata del linguaggio, 17 lingue)
- `xlm-mlm-100-1280` (Modellazione mascherata del linguaggio, 100 lingue)
Questi modelli sono utilizzati per rappresentazioni generiche di frasi, a differenza dei precedenti checkpoints XML.
## BERT
Il seguente modello BERT puรฒ essere usato per compiti multilingue:
- `bert-base-multilingual-uncased` (Modellazione mascherata del linguaggio + Previsione della prossima frase, 102 lingue)
- `bert-base-multilingual-cased` (Modellazione mascherata del linguaggio + Previsione della prossima frase, 104 lingue)
Questi modelli non richiedono language embeddings per fare inferenza. Riescono ad identificare il linguaggio dal contesto e inferire di conseguenza.
## XLM-RoBERTa
Il seguente modello XLM-RoBERTa puรฒ essere usato per compiti multilingue:
- `xlm-roberta-base` (Modellazione mascherata del linguaggio, 100 lingue)
- `xlm-roberta-large` (Modellazione mascherata del linguaggio, 100 lingue)
XLM-RoBERTa รจ stato addestrato su 2.5TB di dati CommonCrawl appena creati e puliti in 100 lingue. Offre notevoli vantaggi rispetto ai modelli multilingue rilasciati in precedenza, come mBERT o XLM, in compiti come la classificazione, l'etichettatura delle sequenze e la risposta alle domande.
## M2M100
Il seguente modello M2M100 puรฒ essere usato per compiti multilingue:
- `facebook/m2m100_418M` (Traduzione)
- `facebook/m2m100_1.2B` (Traduzione)
In questo esempio, carica il checkpoint `facebook/m2m100_418M` per tradurre dal cinese all'inglese. Puoi impostare la lingua di partenza nel tokenizer:
```py
>>> from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer
>>> en_text = "Do not meddle in the affairs of wizards, for they are subtle and quick to anger."
>>> chinese_text = "ไธ่ฆๆๆๅทซๅธซ็ไบๅ, ๅ ็บไปๅๆฏๅพฎๅฆ็, ๅพๅฟซๅฐฑๆ็ผๆ."
>>> tokenizer = M2M100Tokenizer.from_pretrained("facebook/m2m100_418M", src_lang="zh")
>>> model = M2M100ForConditionalGeneration.from_pretrained("facebook/m2m100_418M")
```
Applica il tokenizer al testo:
```py
>>> encoded_zh = tokenizer(chinese_text, return_tensors="pt")
```
M2M100 forza l'id della lingua obiettivo come primo token generato per tradurre nella lingua obiettivo. Imposta il parametro `forced_bos_token_id` a `en` nel metodo `generate` per tradurre in inglese:
```py
>>> generated_tokens = model.generate(**encoded_zh, forced_bos_token_id=tokenizer.get_lang_id("en"))
>>> tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
'Do not interfere with the matters of the witches, because they are delicate and will soon be angry.'
```
## MBart
Il seguente modello MBart puรฒ essere usato per compiti multilingue:
- `facebook/mbart-large-50-one-to-many-mmt` (Traduzione automatica multilingue uno-a-molti, 50 lingue)
- `facebook/mbart-large-50-many-to-many-mmt` (Traduzione automatica multilingue molti-a-molti, 50 lingue)
- `facebook/mbart-large-50-many-to-one-mmt` (Traduzione automatica multilingue molti-a-uno, 50 lingue)
- `facebook/mbart-large-50` (Traduzione multilingue, 50 lingue)
- `facebook/mbart-large-cc25`
In questo esempio, carica il checkpoint `facebook/mbart-large-50-many-to-many-mmt` per tradurre dal finlandese all'inglese. Puoi impostare la lingua di partenza nel tokenizer:
```py
>>> from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
>>> en_text = "Do not meddle in the affairs of wizards, for they are subtle and quick to anger."
>>> fi_text = "รlรค sekaannu velhojen asioihin, sillรค ne ovat hienovaraisia ja nopeasti vihaisia."
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/mbart-large-50-many-to-many-mmt", src_lang="fi_FI")
>>> model = AutoModelForSeq2SeqLM.from_pretrained("facebook/mbart-large-50-many-to-many-mmt")
```
Applica il tokenizer sul testo:
```py
>>> encoded_en = tokenizer(en_text, return_tensors="pt")
```
MBart forza l'id della lingua obiettivo come primo token generato per tradurre nella lingua obiettivo. Imposta il parametro `forced_bos_token_id` a `en` nel metodo `generate` per tradurre in inglese:
```py
>>> generated_tokens = model.generate(**encoded_en, forced_bos_token_id=tokenizer.lang_code_to_id("en_XX"))
>>> tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
"Don't interfere with the wizard's affairs, because they are subtle, will soon get angry."
```
Se stai usando il checkpoint `facebook/mbart-large-50-many-to-one-mmt`, non hai bisogno di forzare l'id della lingua obiettivo come primo token generato altrimenti l'uso รจ lo stesso.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/preprocessing.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Preprocess
[[open-in-colab]]
Prima di poter usare i dati in un modello, bisogna processarli in un formato accettabile per quest'ultimo. Un modello non comprende il testo grezzo, le immagini o l'audio. Bisogna convertire questi input in numeri e assemblarli all'interno di tensori. In questa esercitazione, tu potrai:
* Preprocessare dati testuali con un tokenizer.
* Preprocessare immagini o dati audio con un estrattore di caratteristiche.
* Preprocessare dati per attivitร multimodali mediante un processore.
## NLP
<Youtube id="Yffk5aydLzg"/>
Lo strumento principale per processare dati testuali รจ un [tokenizer](main_classes/tokenizer). Un tokenizer inizia separando il testo in *tokens* secondo una serie di regole. I tokens sono convertiti in numeri, questi vengono utilizzati per costruire i tensori di input del modello. Anche altri input addizionali se richiesti dal modello vengono aggiunti dal tokenizer.
<Tip>
Se stai pensando si utilizzare un modello preaddestrato, รจ importante utilizzare il tokenizer preaddestrato associato. Questo assicura che il testo sia separato allo stesso modo che nel corpus usato per l'addestramento, e venga usata la stessa mappatura tokens-to-index (solitamente indicato come il *vocabolario*) come nel preaddestramento.
</Tip>
Iniziamo subito caricando un tokenizer preaddestrato con la classe [`AutoTokenizer`]. Questo scarica il *vocabolario* usato quando il modello รจ stato preaddestrato.
### Tokenize
Carica un tokenizer preaddestrato con [`AutoTokenizer.from_pretrained`]:
```py
>>> from transformers import AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("bert-base-cased")
```
Poi inserisci le tue frasi nel tokenizer:
```py
>>> encoded_input = tokenizer("Do not meddle in the affairs of wizards, for they are subtle and quick to anger.")
>>> print(encoded_input)
{'input_ids': [101, 2079, 2025, 19960, 10362, 1999, 1996, 3821, 1997, 16657, 1010, 2005, 2027, 2024, 11259, 1998, 4248, 2000, 4963, 1012, 102],
'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}
```
Il tokenizer restituisce un dizionario contenente tre oggetti importanti:
* [input_ids](glossary#input-ids) sono gli indici che corrispondono ad ogni token nella frase.
* [attention_mask](glossary#attention-mask) indicata se un token deve essere elaborato o no.
* [token_type_ids](glossary#token-type-ids) identifica a quale sequenza appartiene un token se รจ presente piรน di una sequenza.
Si possono decodificare gli `input_ids` per farsi restituire l'input originale:
```py
>>> tokenizer.decode(encoded_input["input_ids"])
'[CLS] Do not meddle in the affairs of wizards, for they are subtle and quick to anger. [SEP]'
```
Come si puรฒ vedere, il tokenizer aggiunge due token speciali - `CLS` e `SEP` (classificatore e separatore) - alla frase. Non tutti i modelli hanno bisogno dei token speciali, ma se servono, il tokenizer li aggiungerร automaticamente.
Se ci sono piรน frasi che vuoi processare, passale come una lista al tokenizer:
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_inputs = tokenizer(batch_sentences)
>>> print(encoded_inputs)
{'input_ids': [[101, 1252, 1184, 1164, 1248, 6462, 136, 102],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102]],
'token_type_ids': [[0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]],
'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1]]}
```
### Pad
Questo รจ un argomento importante. Quando processi un insieme di frasi potrebbero non avere tutte la stessa lunghezza. Questo รจ un problema perchรจ i tensori, in input del modello, devono avere dimensioni uniformi. Il padding รจ una strategia per assicurarsi che i tensori siano rettangolari aggiungendo uno speciale *padding token* alle frasi piรน corte.
Imposta il parametro `padding` a `True` per imbottire le frasi piรน corte nel gruppo in modo che combacino con la massima lunghezza presente:
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch_sentences, padding=True)
>>> print(encoded_input)
{'input_ids': [[101, 1252, 1184, 1164, 1248, 6462, 136, 102, 0, 0, 0, 0, 0, 0, 0],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102, 0, 0, 0, 0, 0, 0, 0, 0]],
'token_type_ids': [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]],
'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]]}
```
Nota che il tokenizer aggiunge alle sequenze degli `0` perchรจ sono troppo corte!
### Truncation
L'altra faccia della medaglia รจ che avolte le sequenze possono essere troppo lunghe per essere gestite dal modello. In questo caso, avrai bisogno di troncare la sequenza per avere una lunghezza minore.
Imposta il parametro `truncation` a `True` per troncare una sequenza alla massima lunghezza accettata dal modello:
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch_sentences, padding=True, truncation=True)
>>> print(encoded_input)
{'input_ids': [[101, 1252, 1184, 1164, 1248, 6462, 136, 102, 0, 0, 0, 0, 0, 0, 0],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102, 0, 0, 0, 0, 0, 0, 0, 0]],
'token_type_ids': [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]],
'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]]}
```
### Costruire i tensori
Infine, vuoi che il tokenizer restituisca i tensori prodotti dal modello.
Imposta il parametro `return_tensors` su `pt` per PyTorch, o `tf` per TensorFlow:
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch, padding=True, truncation=True, return_tensors="pt")
>>> print(encoded_input)
{'input_ids': tensor([[ 101, 153, 7719, 21490, 1122, 1114, 9582, 1623, 102],
[ 101, 5226, 1122, 9649, 1199, 2610, 1236, 102, 0]]),
'token_type_ids': tensor([[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0]]),
'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 1, 0]])}
===PT-TF-SPLIT===
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch, padding=True, truncation=True, return_tensors="tf")
>>> print(encoded_input)
{'input_ids': <tf.Tensor: shape=(2, 9), dtype=int32, numpy=
array([[ 101, 153, 7719, 21490, 1122, 1114, 9582, 1623, 102],
[ 101, 5226, 1122, 9649, 1199, 2610, 1236, 102, 0]],
dtype=int32)>,
'token_type_ids': <tf.Tensor: shape=(2, 9), dtype=int32, numpy=
array([[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0]], dtype=int32)>,
'attention_mask': <tf.Tensor: shape=(2, 9), dtype=int32, numpy=
array([[1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 1, 0]], dtype=int32)>}
```
## Audio
Gli input audio sono processati in modo differente rispetto al testo, ma l'obiettivo rimane lo stesso: creare sequenze numeriche che il modello puรฒ capire. Un [estrattore di caratteristiche](main_classes/feature_extractor) รจ progettato con lo scopo preciso di estrarre caratteristiche da immagini o dati audio grezzi e convertirli in tensori. Prima di iniziare, installa ๐ค Datasets per caricare un dataset audio e sperimentare:
```bash
pip install datasets
```
Carica il dataset [MInDS-14](https://huggingface.co/datasets/PolyAI/minds14) (vedi il ๐ค [Datasets tutorial](https://huggingface.co/docs/datasets/load_hub) per avere maggiori dettagli su come caricare un dataset):
```py
>>> from datasets import load_dataset, Audio
>>> dataset = load_dataset("PolyAI/minds14", name="en-US", split="train")
```
Accedi al primo elemento della colonna `audio` per dare uno sguardo all'input. Richiamando la colonna `audio` sarร caricato automaticamente e ricampionato il file audio:
```py
>>> dataset[0]["audio"]
{'array': array([ 0. , 0.00024414, -0.00024414, ..., -0.00024414,
0. , 0. ], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/f14948e0e84be638dd7943ac36518a4cf3324e8b7aa331c5ab11541518e9368c/en-US~JOINT_ACCOUNT/602ba55abb1e6d0fbce92065.wav',
'sampling_rate': 8000}
```
Questo restituisce tre oggetti:
* `array` รจ il segnale vocale caricato - e potenzialmente ricampionato - come vettore 1D.
* `path` il percorso del file audio.
* `sampling_rate` si riferisce al numero di campioni del segnale vocale misurati al secondo.
### Ricampionamento
Per questo tutorial, puoi usare il modello [Wav2Vec2](https://huggingface.co/facebook/wav2vec2-base). Come puoi vedere dalla model card, il modello Wav2Vec2 รจ preaddestrato su un campionamento vocale a 16kHz.ร importante che la frequenza di campionamento dei tuoi dati audio combaci con la frequenza di campionamento del dataset usato per preaddestrare il modello. Se la frequenza di campionamento dei tuoi dati non รจ uguale dovrai ricampionare i tuoi dati audio.
Per esempio, il dataset [MInDS-14](https://huggingface.co/datasets/PolyAI/minds14) ha una frequenza di campionamento di 8000kHz. Utilizzando il modello Wav2Vec2 su questo dataset, alzala a 16kHz:
```py
>>> dataset = load_dataset("PolyAI/minds14", name="en-US", split="train")
>>> dataset[0]["audio"]
{'array': array([ 0. , 0.00024414, -0.00024414, ..., -0.00024414,
0. , 0. ], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/f14948e0e84be638dd7943ac36518a4cf3324e8b7aa331c5ab11541518e9368c/en-US~JOINT_ACCOUNT/602ba55abb1e6d0fbce92065.wav',
'sampling_rate': 8000}
```
1. Usa il metodo di ๐ค Datasets' [`cast_column`](https://huggingface.co/docs/datasets/package_reference/main_classes#datasets.Dataset.cast_column) per alzare la frequenza di campionamento a 16kHz:
```py
>>> dataset = dataset.cast_column("audio", Audio(sampling_rate=16_000))
```
2. Carica il file audio:
```py
>>> dataset[0]["audio"]
{'array': array([ 2.3443763e-05, 2.1729663e-04, 2.2145823e-04, ...,
3.8356509e-05, -7.3497440e-06, -2.1754686e-05], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/f14948e0e84be638dd7943ac36518a4cf3324e8b7aa331c5ab11541518e9368c/en-US~JOINT_ACCOUNT/602ba55abb1e6d0fbce92065.wav',
'sampling_rate': 16000}
```
Come puoi notare, la `sampling_rate` adesso รจ 16kHz!
### Feature extractor
Il prossimo passo รจ caricare un estrattore di caratteristiche per normalizzare e fare padding sull'input. Quando applichiamo il padding sui dati testuali, uno `0` รจ aggiunto alle sequenze piรน brevi. La stessa idea si applica ai dati audio, l'estrattore di caratteristiche per gli audio aggiungerร uno `0` - interpretato come silenzio - agli `array`.
Carica l'estrattore delle caratteristiche con [`AutoFeatureExtractor.from_pretrained`]:
```py
>>> from transformers import AutoFeatureExtractor
>>> feature_extractor = AutoFeatureExtractor.from_pretrained("facebook/wav2vec2-base")
```
Inserisci l' `array` audio nell'estrattore delle caratteristiche. Noi raccomandiamo sempre di aggiungere il parametro `sampling_rate` nell'estrattore delle caratteristiche per correggere meglio qualche errore, dovuto ai silenzi, che potrebbe verificarsi.
```py
>>> audio_input = [dataset[0]["audio"]["array"]]
>>> feature_extractor(audio_input, sampling_rate=16000)
{'input_values': [array([ 3.8106556e-04, 2.7506407e-03, 2.8015103e-03, ...,
5.6335266e-04, 4.6588284e-06, -1.7142107e-04], dtype=float32)]}
```
### Pad e truncate
Come per il tokenizer, puoi applicare le operazioni padding o truncation per manipolare sequenze di variabili a lotti. Dai uno sguaro alla lunghezza delle sequenze di questi due campioni audio:
```py
>>> dataset[0]["audio"]["array"].shape
(173398,)
>>> dataset[1]["audio"]["array"].shape
(106496,)
```
Come puoi vedere, il primo campione ha una sequenza piรน lunga del secondo. Crea una funzione che preprocesserร il dataset. Specifica una lunghezza massima del campione, e l'estrattore di features si occuperร di riempire o troncare la sequenza per coincidervi:
```py
>>> def preprocess_function(examples):
... audio_arrays = [x["array"] for x in examples["audio"]]
... inputs = feature_extractor(
... audio_arrays,
... sampling_rate=16000,
... padding=True,
... max_length=100000,
... truncation=True,
... )
... return inputs
```
Applica la funzione ai primi esempi nel dataset:
```py
>>> processed_dataset = preprocess_function(dataset[:5])
```
Adesso guarda la lunghezza dei campioni elaborati:
```py
>>> processed_dataset["input_values"][0].shape
(100000,)
>>> processed_dataset["input_values"][1].shape
(100000,)
```
La lunghezza dei campioni adesso coincide con la massima lunghezza impostata nelle funzione.
## Vision
Un estrattore di caratteristiche si puรฒ usare anche per processare immagini e per compiti di visione. Ancora una volta, l'obiettivo รจ convertire l'immagine grezza in un lotto di tensori come input.
Carica il dataset [food101](https://huggingface.co/datasets/food101) per questa esercitazione. Usa il parametro `split` di ๐ค Datasets per caricare solo un piccolo campione dal dataset di addestramento poichรจ il set di dati รจ molto grande:
```py
>>> from datasets import load_dataset
>>> dataset = load_dataset("food101", split="train[:100]")
```
Secondo passo, dai uno sguardo alle immagini usando la caratteristica [`Image`](https://huggingface.co/docs/datasets/package_reference/main_classes.html?highlight=image#datasets.Image) di ๐ค Datasets:
```py
>>> dataset[0]["image"]
```
### Feature extractor
Carica l'estrattore di caratteristiche [`AutoFeatureExtractor.from_pretrained`]:
```py
>>> from transformers import AutoFeatureExtractor
>>> feature_extractor = AutoFeatureExtractor.from_pretrained("google/vit-base-patch16-224")
```
### Data augmentation
Per le attivitร di visione, รจ usuale aggiungere alcuni tipi di data augmentation alle immagini come parte del preprocessing. Puoi aggiungere augmentations con qualsiasi libreria che preferisci, ma in questa esercitazione, userai il modulo [`transforms`](https://pytorch.org/vision/stable/transforms.html) di torchvision.
1. Normalizza l'immagine e usa [`Compose`](https://pytorch.org/vision/master/generated/torchvision.transforms.Compose.html) per concatenare alcune trasformazioni - [`RandomResizedCrop`](https://pytorch.org/vision/main/generated/torchvision.transforms.RandomResizedCrop.html) e [`ColorJitter`](https://pytorch.org/vision/main/generated/torchvision.transforms.ColorJitter.html) - insieme:
```py
>>> from torchvision.transforms import Compose, Normalize, RandomResizedCrop, ColorJitter, ToTensor
>>> normalize = Normalize(mean=feature_extractor.image_mean, std=feature_extractor.image_std)
>>> _transforms = Compose(
... [RandomResizedCrop(feature_extractor.size), ColorJitter(brightness=0.5, hue=0.5), ToTensor(), normalize]
... )
```
2. Il modello accetta [`pixel_values`](model_doc/visionencoderdecoder#transformers.VisionEncoderDecoderModel.forward.pixel_values) come input. Questo valore รจ generato dall'estrattore di caratteristiche. Crea una funzione che genera `pixel_values` dai transforms:
```py
>>> def transforms(examples):
... examples["pixel_values"] = [_transforms(image.convert("RGB")) for image in examples["image"]]
... return examples
```
3. Poi utilizza ๐ค Datasets [`set_transform`](https://huggingface.co/docs/datasets/process#format-transform)per applicare al volo la trasformazione:
```py
>>> dataset.set_transform(transforms)
```
4. Adesso quando accedi all'immagine, puoi notare che l'estrattore di caratteristiche ha aggiunto `pixel_values` allo schema di input:
```py
>>> dataset[0]["image"]
{'image': <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=384x512 at 0x7F1A7B0630D0>,
'label': 6,
'pixel_values': tensor([[[ 0.0353, 0.0745, 0.1216, ..., -0.9922, -0.9922, -0.9922],
[-0.0196, 0.0667, 0.1294, ..., -0.9765, -0.9843, -0.9922],
[ 0.0196, 0.0824, 0.1137, ..., -0.9765, -0.9686, -0.8667],
...,
[ 0.0275, 0.0745, 0.0510, ..., -0.1137, -0.1216, -0.0824],
[ 0.0667, 0.0824, 0.0667, ..., -0.0588, -0.0745, -0.0980],
[ 0.0353, 0.0353, 0.0431, ..., -0.0039, -0.0039, -0.0588]],
[[ 0.2078, 0.2471, 0.2863, ..., -0.9451, -0.9373, -0.9451],
[ 0.1608, 0.2471, 0.3098, ..., -0.9373, -0.9451, -0.9373],
[ 0.2078, 0.2706, 0.3020, ..., -0.9608, -0.9373, -0.8275],
...,
[-0.0353, 0.0118, -0.0039, ..., -0.2392, -0.2471, -0.2078],
[ 0.0196, 0.0353, 0.0196, ..., -0.1843, -0.2000, -0.2235],
[-0.0118, -0.0039, -0.0039, ..., -0.0980, -0.0980, -0.1529]],
[[ 0.3961, 0.4431, 0.4980, ..., -0.9216, -0.9137, -0.9216],
[ 0.3569, 0.4510, 0.5216, ..., -0.9059, -0.9137, -0.9137],
[ 0.4118, 0.4745, 0.5216, ..., -0.9137, -0.8902, -0.7804],
...,
[-0.2314, -0.1922, -0.2078, ..., -0.4196, -0.4275, -0.3882],
[-0.1843, -0.1686, -0.2000, ..., -0.3647, -0.3804, -0.4039],
[-0.1922, -0.1922, -0.1922, ..., -0.2941, -0.2863, -0.3412]]])}
```
Di seguito come si vede l'immagine dopo la fase di preprocessing. Come ci si aspetterebbe dalle trasformazioni applicate, l'immagine รจ stata ritagliata in modo casuale e le proprietร del colore sono diverse.
```py
>>> import numpy as np
>>> import matplotlib.pyplot as plt
>>> img = dataset[0]["pixel_values"]
>>> plt.imshow(img.permute(1, 2, 0))
```
## Multimodal
Per attivitร multimodali userai una combinazione di tutto quello che hai imparato poco fa e applicherai le tue competenze alla comprensione automatica del parlato (Automatic Speech Recognition - ASR). Questo significa che avrai bisogno di:
* Un estrattore delle caratteristiche per processare i dati audio.
* Il Tokenizer per processare i testi.
Ritorna sul datasere [LJ Speech](https://huggingface.co/datasets/lj_speech):
```py
>>> from datasets import load_dataset
>>> lj_speech = load_dataset("lj_speech", split="train")
```
Visto che sei interessato solo alle colonne `audio` e `text`, elimina tutte le altre:
```py
>>> lj_speech = lj_speech.map(remove_columns=["file", "id", "normalized_text"])
```
Adesso guarda le colonne `audio` e `text`:
```py
>>> lj_speech[0]["audio"]
{'array': array([-7.3242188e-04, -7.6293945e-04, -6.4086914e-04, ...,
7.3242188e-04, 2.1362305e-04, 6.1035156e-05], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/917ece08c95cf0c4115e45294e3cd0dee724a1165b7fc11798369308a465bd26/LJSpeech-1.1/wavs/LJ001-0001.wav',
'sampling_rate': 22050}
>>> lj_speech[0]["text"]
'Printing, in the only sense with which we are at present concerned, differs from most if not from all the arts and crafts represented in the Exhibition'
```
Ricorda dalla sezione precedente sull'elaborazione dei dati audio, tu dovresti sempre [ricampionare](preprocessing#audio) la frequenza di campionamento dei tuoi dati audio per farla coincidere con quella del dataset usato dal modello preaddestrato:
```py
>>> lj_speech = lj_speech.cast_column("audio", Audio(sampling_rate=16_000))
```
### Processor
Un processor combina un estrattore di caratteristiche e un tokenizer. Carica un processor con [`AutoProcessor.from_pretrained]:
```py
>>> from transformers import AutoProcessor
>>> processor = AutoProcessor.from_pretrained("facebook/wav2vec2-base-960h")
```
1. Crea una funzione che processi i dati audio in `input_values`, e tokenizza il testo in `labels`. Questi sono i tuoi input per il modello:
```py
>>> def prepare_dataset(example):
... audio = example["audio"]
... example.update(processor(audio=audio["array"], text=example["text"], sampling_rate=16000))
... return example
```
2. Applica la funzione `prepare_dataset` ad un campione:
```py
>>> prepare_dataset(lj_speech[0])
```
Nota che il processor ha aggiunto `input_values` e `labels`. La frequenza di campionamento รจ stata corretta riducendola a 16kHz.
Fantastico, ora dovresti essere in grado di preelaborare i dati per qualsiasi modalitร e persino di combinare modalitร diverse! Nella prossima esercitazione, impareremo a mettere a punto un modello sui dati appena pre-elaborati.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/model_sharing.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Condividi un modello
Gli ultimi due tutorial ti hanno mostrato come puoi fare fine-tuning di un modello con PyTorch, Keras e ๐ค Accelerate per configurazioni distribuite. Il prossimo passo รจ quello di condividere il tuo modello con la community! In Hugging Face, crediamo nella condivisione della conoscenza e delle risorse in modo da democratizzare l'intelligenza artificiale per chiunque. Ti incoraggiamo a considerare di condividere il tuo modello con la community per aiutare altre persone a risparmiare tempo e risorse.
In questo tutorial, imparerai due metodi per la condivisione di un modello trained o fine-tuned nel [Model Hub](https://huggingface.co/models):
- Condividi in modo programmatico i tuoi file nell'Hub.
- Trascina i tuoi file nell'Hub mediante interfaccia grafica.
<iframe width="560" height="315" src="https://www.youtube.com/embed/XvSGPZFEjDY" title="YouTube video player"
frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope;
picture-in-picture" allowfullscreen></iframe>
<Tip>
Per condividere un modello con la community, hai bisogno di un account su [huggingface.co](https://huggingface.co/join). Puoi anche unirti ad un'organizzazione esistente o crearne una nuova.
</Tip>
## Caratteristiche dei repository
Ogni repository nel Model Hub si comporta come un tipico repository di GitHub. I nostri repository offrono il versionamento, la cronologia dei commit, e la possibilitร di visualizzare le differenze.
Il versionamento all'interno del Model Hub รจ basato su git e [git-lfs](https://git-lfs.github.com/). In altre parole, puoi trattare un modello come un unico repository, consentendo un maggiore controllo degli accessi e maggiore scalabilitร . Il controllo delle versioni consente *revisions*, un metodo per appuntare una versione specifica di un modello con un hash di commit, un tag o un branch.
Come risultato, puoi caricare una specifica versione di un modello con il parametro `revision`:
```py
>>> model = AutoModel.from_pretrained(
... "julien-c/EsperBERTo-small", revision="v2.0.1" # nome di un tag, di un branch, o commit hash
... )
```
Anche i file possono essere modificati facilmente in un repository ed รจ possibile visualizzare la cronologia dei commit e le differenze:
## Configurazione
Prima di condividere un modello nell'Hub, hai bisogno delle tue credenziali di Hugging Face. Se hai accesso ad un terminale, esegui il seguente comando nell'ambiente virtuale in cui รจ installata la libreria ๐ค Transformers. Questo memorizzerร il tuo token di accesso nella cartella cache di Hugging Face (di default `~/.cache/`):
```bash
huggingface-cli login
```
Se stai usando un notebook come Jupyter o Colaboratory, assicurati di avere la libreria [`huggingface_hub`](https://huggingface.co/docs/hub/adding-a-library) installata. Questa libreria ti permette di interagire in maniera programmatica con l'Hub.
```bash
pip install huggingface_hub
```
Utilizza `notebook_login` per accedere all'Hub, e segui il link [qui](https://huggingface.co/settings/token) per generare un token con cui effettuare il login:
```py
>>> from huggingface_hub import notebook_login
>>> notebook_login()
```
## Converti un modello per tutti i framework
Per assicurarti che il tuo modello possa essere utilizzato da persone che lavorano con un framework differente, ti raccomandiamo di convertire e caricare il tuo modello sia con i checkpoint di PyTorch che con quelli di TensorFlow. Anche se รจ possibile caricare il modello da un framework diverso, se si salta questo passaggio, il caricamento sarร piรน lento perchรฉ ๐ค Transformers ha bisogno di convertire i checkpoint al momento.
Convertire un checkpoint per un altro framework รจ semplice. Assicurati di avere PyTorch e TensorFlow installati (vedi [qui](installation) per le istruzioni d'installazione), e poi trova il modello specifico per il tuo compito nell'altro framework.
<frameworkcontent>
<pt>
Specifica `from_tf=True` per convertire un checkpoint da TensorFlow a PyTorch:
```py
>>> pt_model = DistilBertForSequenceClassification.from_pretrained(
... "path/verso/il-nome-magnifico-che-hai-scelto", from_tf=True
... )
>>> pt_model.save_pretrained("path/verso/il-nome-magnifico-che-hai-scelto")
```
</pt>
<tf>
Specifica `from_pt=True` per convertire un checkpoint da PyTorch a TensorFlow:
```py
>>> tf_model = TFDistilBertForSequenceClassification.from_pretrained(
... "path/verso/il-nome-magnifico-che-hai-scelto", from_pt=True
... )
```
Poi puoi salvare il tuo nuovo modello in TensorFlow con il suo nuovo checkpoint:
```py
>>> tf_model.save_pretrained("path/verso/il-nome-magnifico-che-hai-scelto")
```
</tf>
<jax>
Se un modello รจ disponibile in Flax, puoi anche convertire un checkpoint da PyTorch a Flax:
```py
>>> flax_model = FlaxDistilBertForSequenceClassification.from_pretrained(
... "path/verso/il-nome-magnifico-che-hai-scelto", from_pt=True
... )
```
</jax>
</frameworkcontent>
## Condividi un modello durante il training
<frameworkcontent>
<pt>
<Youtube id="Z1-XMy-GNLQ"/>
Condividere un modello nell'Hub รจ tanto semplice quanto aggiungere un parametro extra o un callback. Ricorda dal [tutorial sul fine-tuning](training), la classe [`TrainingArguments`] รจ dove specifichi gli iperparametri e le opzioni addizionali per l'allenamento. Una di queste opzioni di training include l'abilitร di condividere direttamente un modello nell'Hub. Imposta `push_to_hub=True` in [`TrainingArguments`]:
```py
>>> training_args = TrainingArguments(output_dir="il-mio-bellissimo-modello", push_to_hub=True)
```
Passa gli argomenti per il training come di consueto al [`Trainer`]:
```py
>>> trainer = Trainer(
... model=model,
... args=training_args,
... train_dataset=small_train_dataset,
... eval_dataset=small_eval_dataset,
... compute_metrics=compute_metrics,
... )
```
Dopo aver effettuato il fine-tuning del tuo modello, chiama [`~transformers.Trainer.push_to_hub`] sul [`Trainer`] per condividere il modello allenato nell'Hub. ๐ค Transformers aggiungerร in modo automatico persino gli iperparametri, i risultati del training e le versioni del framework alla scheda del tuo modello (model card, in inglese)!
```py
>>> trainer.push_to_hub()
```
</pt>
<tf>
Condividi un modello nell'Hub con [`PushToHubCallback`]. Nella funzione [`PushToHubCallback`], aggiungi:
- Una directory di output per il tuo modello.
- Un tokenizer.
- L'`hub_model_id`, che รจ il tuo username sull'Hub e il nome del modello.
```py
>>> from transformers import PushToHubCallback
>>> push_to_hub_callback = PushToHubCallback(
... output_dir="./il_path_dove_salvare_il_tuo_modello",
... tokenizer=tokenizer,
... hub_model_id="il-tuo-username/il-mio-bellissimo-modello",
... )
```
Aggiungi il callback a [`fit`](https://keras.io/api/models/model_training_apis/), e ๐ค Transformers caricherร il modello allenato nell'Hub:
```py
>>> model.fit(tf_train_dataset, validation_data=tf_validation_dataset, epochs=3, callbacks=push_to_hub_callback)
```
</tf>
</frameworkcontent>
## Utilizzare la funzione `push_to_hub`
Puoi anche chiamare `push_to_hub` direttamente sul tuo modello per caricarlo nell'Hub.
Specifica il nome del tuo modello in `push_to_hub`:
```py
>>> pt_model.push_to_hub("il-mio-bellissimo-modello")
```
Questo crea un repository sotto il proprio username con il nome del modello `il-mio-bellissimo-modello`. Ora chiunque puรฒ caricare il tuo modello con la funzione `from_pretrained`:
```py
>>> from transformers import AutoModel
>>> model = AutoModel.from_pretrained("il-tuo-username/il-mio-bellissimo-modello")
```
Se fai parte di un'organizzazione e vuoi invece condividere un modello sotto il nome dell'organizzazione, aggiungi il parametro `organization`:
```py
>>> pt_model.push_to_hub("il-mio-bellissimo-modello", organization="la-mia-fantastica-org")
```
La funzione `push_to_hub` puรฒ essere anche utilizzata per aggiungere altri file al repository del modello. Per esempio, aggiungi un tokenizer ad un repository di un modello:
```py
>>> tokenizer.push_to_hub("il-mio-bellissimo-modello")
```
O magari potresti voler aggiungere la versione di TensorFlow del tuo modello PyTorch a cui hai fatto fine-tuning:
```py
>>> tf_model.push_to_hub("il-mio-bellissimo-modello")
```
Ora quando navighi nel tuo profilo Hugging Face, dovresti vedere il tuo repository del modello appena creato. Premendo sulla scheda **Files** vengono visualizzati tutti i file caricati nel repository.
Per maggiori dettagli su come creare e caricare file ad un repository, fai riferimento alla documentazione [qui](https://huggingface.co/docs/hub/how-to-upstream).
## Carica un modello utilizzando l'interfaccia web
Chi preferisce un approccio senza codice puรฒ caricare un modello tramite l'interfaccia web dell'hub. Visita [huggingface.co/new](https://huggingface.co/new) per creare un nuovo repository:
Da qui, aggiungi alcune informazioni sul tuo modello:
- Seleziona il/la **owner** del repository. Puoi essere te o qualunque organizzazione di cui fai parte.
- Scegli un nome per il tuo modello, il quale sarร anche il nome del repository.
- Scegli se il tuo modello รจ pubblico o privato.
- Specifica la licenza utilizzata per il tuo modello.
Ora premi sulla scheda **Files** e premi sul pulsante **Add file** per caricare un nuovo file al tuo repository. Trascina poi un file per caricarlo e aggiungere un messaggio di commit.
## Aggiungi una scheda del modello
Per assicurarti che chiunque possa comprendere le abilitร , limitazioni, i potenziali bias e le considerazioni etiche del tuo modello, per favore aggiungi una scheda del modello (model card, in inglese) al tuo repository. La scheda del modello รจ definita nel file `README.md`. Puoi aggiungere una scheda del modello:
* Creando manualmente e caricando un file `README.md`.
* Premendo sul pulsante **Edit model card** nel repository del tuo modello.
Dai un'occhiata alla [scheda del modello](https://huggingface.co/distilbert-base-uncased) di DistilBert per avere un buon esempio del tipo di informazioni che una scheda di un modello deve includere. Per maggiori dettagli legati ad altre opzioni che puoi controllare nel file `README.md`, come l'impatto ambientale o widget di esempio, fai riferimento alla documentazione [qui](https://huggingface.co/docs/hub/models-cards).
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/create_a_model.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Crea un'architettura personalizzata
Una [`AutoClass`](model_doc/auto) deduce automaticamente il modello dell'architettura e scarica la configurazione e i pesi pre-allenati. Generalmente, noi consigliamo di usare un `AutoClass` per produrre un codice indipendente dal checkpoint. Ma gli utenti che desiderano un controllo maggiore su parametri specifici del modello possono creare un modello ๐ค Transformers personalizzato da poche classi base. Questo potrebbe essere particolarmente utile per qualunque persona sia interessata nel studiare, allenare o sperimentare con un modello ๐ค Transformers. In questa guida, approfondisci la creazione di un modello personalizzato senza `AutoClass`. Impara come:
- Caricare e personalizzare una configurazione del modello.
- Creare un'architettura modello.
- Creare un tokenizer lento e veloce per il testo.
- Creare un estrattore di caratteristiche per attivitร riguardanti audio o immagini.
- Creare un processore per attivitร multimodali.
## Configurazione
Una [configurazione](main_classes/configuration) si riferisce agli attributi specifici di un modello. Ogni configurazione del modello ha attributi diversi; per esempio, tutti i modelli npl hanno questi attributi in comune `hidden_size`, `num_attention_heads`, `num_hidden_layers` e `vocab_size`. Questi attributi specificano il numero di attention heads o strati nascosti con cui costruire un modello.
Dai un'occhiata piรน da vicino a [DistilBERT](model_doc/distilbert) accedendo a [`DistilBertConfig`] per ispezionare i suoi attributi:
```py
>>> from transformers import DistilBertConfig
>>> config = DistilBertConfig()
>>> print(config)
DistilBertConfig {
"activation": "gelu",
"attention_dropout": 0.1,
"dim": 768,
"dropout": 0.1,
"hidden_dim": 3072,
"initializer_range": 0.02,
"max_position_embeddings": 512,
"model_type": "distilbert",
"n_heads": 12,
"n_layers": 6,
"pad_token_id": 0,
"qa_dropout": 0.1,
"seq_classif_dropout": 0.2,
"sinusoidal_pos_embds": false,
"transformers_version": "4.16.2",
"vocab_size": 30522
}
```
[`DistilBertConfig`] mostra tutti gli attributi predefiniti usati per costruire una base [`DistilBertModel`]. Tutti gli attributi sono personalizzabili, creando uno spazio per sperimentare. Per esempio, puoi configurare un modello predefinito per:
- Provare un funzione di attivazione diversa con il parametro `activation`.
- Utilizzare tasso di drop out piรน elevato per le probalitร di attention con il parametro `attention_dropout`.
```py
>>> my_config = DistilBertConfig(activation="relu", attention_dropout=0.4)
>>> print(my_config)
DistilBertConfig {
"activation": "relu",
"attention_dropout": 0.4,
"dim": 768,
"dropout": 0.1,
"hidden_dim": 3072,
"initializer_range": 0.02,
"max_position_embeddings": 512,
"model_type": "distilbert",
"n_heads": 12,
"n_layers": 6,
"pad_token_id": 0,
"qa_dropout": 0.1,
"seq_classif_dropout": 0.2,
"sinusoidal_pos_embds": false,
"transformers_version": "4.16.2",
"vocab_size": 30522
}
```
Nella funzione [`~PretrainedConfig.from_pretrained`] possono essere modificati gli attributi del modello pre-allenato:
```py
>>> my_config = DistilBertConfig.from_pretrained("distilbert-base-uncased", activation="relu", attention_dropout=0.4)
```
Quando la configurazione del modello ti soddisfa, la puoi salvare con [`~PretrainedConfig.save_pretrained`]. Il file della tua configurazione รจ memorizzato come file JSON nella save directory specificata:
```py
>>> my_config.save_pretrained(save_directory="./your_model_save_path")
```
Per riutilizzare la configurazione del file, caricalo con [`~PretrainedConfig.from_pretrained`]:
```py
>>> my_config = DistilBertConfig.from_pretrained("./your_model_save_path/my_config.json")
```
<Tip>
Puoi anche salvare il file di configurazione come dizionario oppure come la differenza tra gli attributi della tua configurazione personalizzata e gli attributi della configurazione predefinita! Guarda la documentazione [configuration](main_classes/configuration) per piรน dettagli.
</Tip>
## Modello
Il prossimo passo e di creare [modello](main_classes/models). Il modello - vagamente riferito anche come architettura - definisce cosa ogni strato deve fare e quali operazioni stanno succedendo. Attributi come `num_hidden_layers` provenienti dalla configurazione sono usati per definire l'architettura. Ogni modello condivide la classe base [`PreTrainedModel`] e alcuni metodi comuni come il ridimensionamento degli input embeddings e la soppressione delle self-attention heads . Inoltre, tutti i modelli sono la sottoclasse di [`torch.nn.Module`](https://pytorch.org/docs/stable/generated/torch.nn.Module.html), [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) o [`flax.linen.Module`](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html). Cio significa che i modelli sono compatibili con l'uso di ciascun di framework.
<frameworkcontent>
<pt>
Carica gli attributi della tua configurazione personalizzata nel modello:
```py
>>> from transformers import DistilBertModel
>>> my_config = DistilBertConfig.from_pretrained("./your_model_save_path/my_config.json")
>>> model = DistilBertModel(my_config)
```
Questo crea modelli con valori casuali invece di pesi pre-allenati. Non sarai in grado di usare questo modello per niente di utile finchรฉ non lo alleni. L'allenamento รจ un processo costoso e che richiede tempo . Generalmente รจ meglio usare un modello pre-allenato per ottenere risultati migliori velocemente, utilizzando solo una frazione delle risorse neccesarie per l'allenamento.
Crea un modello pre-allenato con [`~PreTrainedModel.from_pretrained`]:
```py
>>> model = DistilBertModel.from_pretrained("distilbert-base-uncased")
```
Quando carichi pesi pre-allenati, la configurazione del modello predefinito รจ automaticamente caricata se il modello รจ fornito da ๐ค Transformers. Tuttavia, puoi ancora sostituire gli attributi - alcuni o tutti - di configurazione del modello predefinito con i tuoi se lo desideri:
```py
>>> model = DistilBertModel.from_pretrained("distilbert-base-uncased", config=my_config)
```
</pt>
<tf>
Carica gli attributi di configurazione personalizzati nel modello:
```py
>>> from transformers import TFDistilBertModel
>>> my_config = DistilBertConfig.from_pretrained("./your_model_save_path/my_config.json")
>>> tf_model = TFDistilBertModel(my_config)
```
Questo crea modelli con valori casuali invece di pesi pre-allenati. Non sarai in grado di usare questo modello per niente di utile finchรฉ non lo alleni. L'allenamento รจ un processo costoso e che richiede tempo . Generalmente รจ meglio usare un modello pre-allenato per ottenere risultati migliori velocemente, utilizzando solo una frazione delle risorse neccesarie per l'allenamento.
Crea un modello pre-allenoto con [`~TFPreTrainedModel.from_pretrained`]:
```py
>>> tf_model = TFDistilBertModel.from_pretrained("distilbert-base-uncased")
```
Quando carichi pesi pre-allenati, la configurazione del modello predefinito รจ automaticamente caricato se il modello รจ fornito da ๐ค Transformers. Tuttavia, puoi ancora sostituire gli attributi - alcuni o tutti - di configurazione del modello predefinito con i tuoi se lo desideri:
```py
>>> tf_model = TFDistilBertModel.from_pretrained("distilbert-base-uncased", config=my_config)
```
</tf>
</frameworkcontent>
### Model head
A questo punto, hai un modello DistilBERT base i cui output sono gli *hidden states* (in italiano stati nascosti). Gli stati nascosti sono passati come input a un model head per produrre l'output finale. ๐ค Transformers fornisce un model head diverso per ogni attivitร fintanto che il modello supporta l'attivitร (i.e., non puoi usare DistilBERT per un attivitร sequence-to-sequence come la traduzione).
<frameworkcontent>
<pt>
Per esempio, [`DistilBertForSequenceClassification`] รจ un modello DistilBERT base con una testa di classificazione per sequenze. La sequenza di classificazione head รจ uno strato lineare sopra gli output ragruppati.
```py
>>> from transformers import DistilBertForSequenceClassification
>>> model = DistilBertForSequenceClassification.from_pretrained("distilbert-base-uncased")
```
Riutilizza facilmente questo checkpoint per un'altra attivitร passando ad un model head differente. Per un attivitร di risposta alle domande, utilizzerai il model head [`DistilBertForQuestionAnswering`]. La head per compiti di question answering รจ simile alla classificazione di sequenza head tranne per il fatto che รจ uno strato lineare sopra l'output degli stati nascosti (hidden states in inglese)
```py
>>> from transformers import DistilBertForQuestionAnswering
>>> model = DistilBertForQuestionAnswering.from_pretrained("distilbert-base-uncased")
```
</pt>
<tf>
Per esempio, [`TFDistilBertForSequenceClassification`] รจ un modello DistilBERT base con classificazione di sequenza head. La classificazione di sequenza head รจ uno strato lineare sopra gli output raggruppati.
```py
>>> from transformers import TFDistilBertForSequenceClassification
>>> tf_model = TFDistilBertForSequenceClassification.from_pretrained("distilbert-base-uncased")
```
Riutilizza facilmente questo checkpoint per un altra attivitร passando ad un modello head diverso. Per un attivitร di risposta alle domande, utilizzerai il model head [`TFDistilBertForQuestionAnswering`]. Il head di risposta alle domande รจ simile alla sequenza di classificazione head tranne per il fatto che รจ uno strato lineare sopra l'output degli stati nascosti (hidden states in inglese)
```py
>>> from transformers import TFDistilBertForQuestionAnswering
>>> tf_model = TFDistilBertForQuestionAnswering.from_pretrained("distilbert-base-uncased")
```
</tf>
</frameworkcontent>
## Tokenizer
L'ultima classe base di cui hai bisogno prima di utilizzare un modello per i dati testuali รจ un [tokenizer](main_classes/tokenizer) per convertire il testo grezzo in tensori. Ci sono due tipi di tokenizer che puoi usare con ๐ค Transformers:
- [`PreTrainedTokenizer`]: un'implementazione Python di un tokenizer.
- [`PreTrainedTokenizerFast`]: un tokenizer dalla nostra libreria [๐ค Tokenizer](https://huggingface.co/docs/tokenizers/python/latest/) basata su Rust. Questo tipo di tokenizer รจ significativamente piรน veloce, specialmente durante la batch tokenization, grazie alla sua implementazione Rust. Il tokenizer veloce offre anche metodi aggiuntivi come *offset mapping* che associa i token alle loro parole o caratteri originali.
Entrambi i tokenizer supportano metodi comuni come la codifica e la decodifica, l'aggiunta di nuovi token e la gestione di token speciali.
<Tip warning={true}>
Non tutti i modelli supportano un tokenizer veloce. Dai un'occhiata a questo [tabella](index#supported-frameworks) per verificare se un modello ha il supporto per tokenizer veloce.
</Tip>
Se hai addestrato il tuo tokenizer, puoi crearne uno dal tuo file *vocabolario*:
```py
>>> from transformers import DistilBertTokenizer
>>> my_tokenizer = DistilBertTokenizer(vocab_file="my_vocab_file.txt", do_lower_case=False, padding_side="left")
```
ร importante ricordare che il vocabolario di un tokenizer personalizzato sarร diverso dal vocabolario generato dal tokenizer di un modello preallenato. ร necessario utilizzare il vocabolario di un modello preallenato se si utilizza un modello preallenato, altrimenti gli input non avranno senso. Crea un tokenizer con il vocabolario di un modello preallenato con la classe [`DistilBertTokenizer`]:
```py
>>> from transformers import DistilBertTokenizer
>>> slow_tokenizer = DistilBertTokenizer.from_pretrained("distilbert-base-uncased")
```
Crea un tokenizer veloce con la classe [`DistilBertTokenizerFast`]:
```py
>>> from transformers import DistilBertTokenizerFast
>>> fast_tokenizer = DistilBertTokenizerFast.from_pretrained("distilbert-base-uncased")
```
<Tip>
Per l'impostazione predefinita, [`AutoTokenizer`] proverร a caricare un tokenizer veloce. Puoi disabilitare questo comportamento impostando `use_fast=False` in `from_pretrained`.
</Tip>
## Estrattore Di Feature
Un estrattore di caratteristiche (feature in inglese) elabora input audio o immagini. Eredita dalla classe [`~feature_extraction_utils.FeatureExtractionMixin`] base e puรฒ anche ereditare dalla classe [`ImageFeatureExtractionMixin`] per l'elaborazione delle caratteristiche dell'immagine o dalla classe [`SequenceFeatureExtractor`] per l'elaborazione degli input audio.
A seconda che tu stia lavorando a un'attivitร audio o visiva, crea un estrattore di caratteristiche associato al modello che stai utilizzando. Ad esempio, crea un [`ViTFeatureExtractor`] predefinito se stai usando [ViT](model_doc/vit) per la classificazione delle immagini:
```py
>>> from transformers import ViTFeatureExtractor
>>> vit_extractor = ViTFeatureExtractor()
>>> print(vit_extractor)
ViTFeatureExtractor {
"do_normalize": true,
"do_resize": true,
"feature_extractor_type": "ViTFeatureExtractor",
"image_mean": [
0.5,
0.5,
0.5
],
"image_std": [
0.5,
0.5,
0.5
],
"resample": 2,
"size": 224
}
```
<Tip>
Se non stai cercando alcuna personalizzazione, usa il metodo `from_pretrained` per caricare i parametri di default dell'estrattore di caratteristiche di un modello.
</Tip>
Modifica uno qualsiasi dei parametri [`ViTFeatureExtractor`] per creare il tuo estrattore di caratteristiche personalizzato:
```py
>>> from transformers import ViTFeatureExtractor
>>> my_vit_extractor = ViTFeatureExtractor(resample="PIL.Image.BOX", do_normalize=False, image_mean=[0.3, 0.3, 0.3])
>>> print(my_vit_extractor)
ViTFeatureExtractor {
"do_normalize": false,
"do_resize": true,
"feature_extractor_type": "ViTFeatureExtractor",
"image_mean": [
0.3,
0.3,
0.3
],
"image_std": [
0.5,
0.5,
0.5
],
"resample": "PIL.Image.BOX",
"size": 224
}
```
Per gli input audio, puoi creare un [`Wav2Vec2FeatureExtractor`] e personalizzare i parametri in modo simile:
```py
>>> from transformers import Wav2Vec2FeatureExtractor
>>> w2v2_extractor = Wav2Vec2FeatureExtractor()
>>> print(w2v2_extractor)
Wav2Vec2FeatureExtractor {
"do_normalize": true,
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
"feature_size": 1,
"padding_side": "right",
"padding_value": 0.0,
"return_attention_mask": false,
"sampling_rate": 16000
}
```
## Processore
Per modelli che supportano attivitร multimodali, ๐ค Transformers offre una classe di processore che racchiude comodamente un estrattore di caratteristiche e un tokenizer in un unico oggetto. Ad esempio, utilizziamo [`Wav2Vec2Processor`] per un'attivitร di riconoscimento vocale automatico (ASR). ASR trascrive l'audio in testo, quindi avrai bisogno di un estrattore di caratteristiche e di un tokenizer.
Crea un estrattore di feature per gestire gli input audio:
```py
>>> from transformers import Wav2Vec2FeatureExtractor
>>> feature_extractor = Wav2Vec2FeatureExtractor(padding_value=1.0, do_normalize=True)
```
Crea un tokenizer per gestire gli input di testo:
```py
>>> from transformers import Wav2Vec2CTCTokenizer
>>> tokenizer = Wav2Vec2CTCTokenizer(vocab_file="my_vocab_file.txt")
```
Combinare l'estrattore di caratteristiche e il tokenizer in [`Wav2Vec2Processor`]:
```py
>>> from transformers import Wav2Vec2Processor
>>> processor = Wav2Vec2Processor(feature_extractor=feature_extractor, tokenizer=tokenizer)
```
Con due classi di base - configurazione e modello - e una classe di preelaborazione aggiuntiva (tokenizer, estrattore di caratteristiche o processore), puoi creare qualsiasi modello supportato da ๐ค Transformers. Ognuna di queste classi base รจ configurabile, consentendoti di utilizzare gli attributi specifici che desideri. ร possibile impostare facilmente un modello per l'addestramento o modificare un modello preallenato esistente per la messa a punto.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/installation.md
|
<!---
Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Installazione
Installa ๐ค Transformers per qualsiasi libreria di deep learning con cui stai lavorando, imposta la tua cache, e opzionalmente configura ๐ค Transformers per l'esecuzione offline.
๐ค Transformers รจ testato su Python 3.6+, PyTorch 1.1.0+, TensorFlow 2.0+, e Flax. Segui le istruzioni di installazione seguenti per la libreria di deep learning che stai utilizzando:
* [PyTorch](https://pytorch.org/get-started/locally/) istruzioni di installazione.
* [TensorFlow 2.0](https://www.tensorflow.org/install/pip) istruzioni di installazione.
* [Flax](https://flax.readthedocs.io/en/latest/) istruzioni di installazione.
## Installazione con pip
Puoi installare ๐ค Transformers in un [ambiente virtuale](https://docs.python.org/3/library/venv.html). Se non sei familiare con gli ambienti virtuali in Python, dai un'occhiata a questa [guida](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/). Un ambiente virtuale rende piรน semplice la gestione di progetti differenti, evitando problemi di compatibilitร tra dipendenze.
Inizia creando un ambiente virtuale nella directory del tuo progetto:
```bash
python -m venv .env
```
Attiva l'ambiente virtuale:
```bash
source .env/bin/activate
```
Ora puoi procedere con l'installazione di ๐ค Transformers eseguendo il comando seguente:
```bash
pip install transformers
```
Per il solo supporto della CPU, puoi installare facilmente ๐ค Transformers e una libreria di deep learning in solo una riga. Ad esempio, installiamo ๐ค Transformers e PyTorch con:
```bash
pip install transformers[torch]
```
๐ค Transformers e TensorFlow 2.0:
```bash
pip install transformers[tf-cpu]
```
๐ค Transformers e Flax:
```bash
pip install transformers[flax]
```
Infine, verifica se ๐ค Transformers รจ stato installato in modo appropriato eseguendo il seguente comando. Questo scaricherร un modello pre-allenato:
```bash
python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('we love you'))"
```
Dopodichรฉ stampa l'etichetta e il punteggio:
```bash
[{'label': 'POSITIVE', 'score': 0.9998704791069031}]
```
## Installazione dalla fonte
Installa ๐ค Transformers dalla fonte con il seguente comando:
```bash
pip install git+https://github.com/huggingface/transformers
```
Questo comando installa la versione `main` piรน attuale invece dell'ultima versione stabile. Questo รจ utile per stare al passo con gli ultimi sviluppi. Ad esempio, se un bug รจ stato sistemato da quando รจ uscita l'ultima versione ufficiale ma non รจ stata ancora rilasciata una nuova versione. Tuttavia, questo significa che questa versione `main` puรฒ non essere sempre stabile. Ci sforziamo per mantenere la versione `main` operativa, e la maggior parte dei problemi viene risolta in poche ore o in un giorno. Se riscontri un problema, per favore apri una [Issue](https://github.com/huggingface/transformers/issues) cosรฌ possiamo sistemarlo ancora piรน velocemente!
Controlla se ๐ค Transformers รจ stata installata in modo appropriato con il seguente comando:
```bash
python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('I love you'))"
```
## Installazione modificabile
Hai bisogno di un'installazione modificabile se vuoi:
* Usare la versione `main` del codice dalla fonte.
* Contribuire a ๐ค Transformers e hai bisogno di testare i cambiamenti nel codice.
Clona il repository e installa ๐ค Transformers con i seguenti comandi:
```bash
git clone https://github.com/huggingface/transformers.git
cd transformers
pip install -e .
```
Questi comandi collegheranno la cartella in cui รจ stato clonato il repository e i path delle librerie Python. Python guarderร ora all'interno della cartella clonata, oltre ai normali path delle librerie. Per esempio, se i tuoi pacchetti Python sono installati tipicamente in `~/anaconda3/envs/main/lib/python3.7/site-packages/`, Python cercherร anche nella cartella clonata: `~/transformers/`.
<Tip warning={true}>
Devi tenere la cartella `transformers` se vuoi continuare ad utilizzare la libreria.
</Tip>
Ora puoi facilmente aggiornare il tuo clone all'ultima versione di ๐ค Transformers con il seguente comando:
```bash
cd ~/transformers/
git pull
```
Il tuo ambiente Python troverร la versione `main` di ๐ค Transformers alla prossima esecuzione.
## Installazione con conda
Installazione dal canale conda `huggingface`:
```bash
conda install -c huggingface transformers
```
## Impostazione della cache
I modelli pre-allenati sono scaricati e memorizzati localmente nella cache in: `~/.cache/huggingface/transformers/`. Questa รจ la directory di default data dalla variabile d'ambiente della shell `TRANSFORMERS_CACHE`. Su Windows, la directory di default รจ data da `C:\Users\username\.cache\huggingface\transformers`. Puoi cambiare le variabili d'ambiente della shell indicate in seguito, in ordine di prioritร , per specificare una directory differente per la cache:
1. Variabile d'ambiente della shell (default): `TRANSFORMERS_CACHE`.
2. Variabile d'ambiente della shell: `HF_HOME` + `transformers/`.
3. Variabile d'ambiente della shell: `XDG_CACHE_HOME` + `/huggingface/transformers`.
<Tip>
๐ค Transformers utilizzerร le variabili d'ambiente della shell `PYTORCH_TRANSFORMERS_CACHE` o `PYTORCH_PRETRAINED_BERT_CACHE` se si proviene da un'iterazione precedente di questa libreria e sono state impostate queste variabili d'ambiente, a meno che non si specifichi la variabile d'ambiente della shell `TRANSFORMERS_CACHE`.
</Tip>
## Modalitร Offline
๐ค Transformers puรฒ essere eseguita in un ambiente firewalled o offline utilizzando solo file locali. Imposta la variabile d'ambiente `TRANSFORMERS_OFFLINE=1` per abilitare questo comportamento.
<Tip>
Aggiungi [๐ค Datasets](https://huggingface.co/docs/datasets/) al tuo flusso di lavoro offline di training impostando la variabile d'ambiente `HF_DATASETS_OFFLINE=1`.
</Tip>
Ad esempio, in genere si esegue un programma su una rete normale, protetta da firewall per le istanze esterne, con il seguente comando:
```bash
python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
```
Esegui lo stesso programma in un'istanza offline con:
```bash
HF_DATASETS_OFFLINE=1 TRANSFORMERS_OFFLINE=1 \
python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
```
Lo script viene ora eseguito senza bloccarsi o attendere il timeout, perchรฉ sa di dover cercare solo file locali.
### Ottenere modelli e tokenizer per l'uso offline
Un'altra opzione per utilizzare offline ๐ค Transformers รจ scaricare i file in anticipo, e poi puntare al loro path locale quando hai la necessitร di utilizzarli offline. Ci sono tre modi per fare questo:
* Scarica un file tramite l'interfaccia utente sul [Model Hub](https://huggingface.co/models) premendo sull'icona โ.
* Utilizza il flusso [`PreTrainedModel.from_pretrained`] e [`PreTrainedModel.save_pretrained`]:
1. Scarica i tuoi file in anticipo con [`PreTrainedModel.from_pretrained`]:
```py
>>> from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
>>> tokenizer = AutoTokenizer.from_pretrained("bigscience/T0_3B")
>>> model = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0_3B")
```
2. Salva i tuoi file in una directory specificata con [`PreTrainedModel.save_pretrained`]:
```py
>>> tokenizer.save_pretrained("./il/tuo/path/bigscience_t0")
>>> model.save_pretrained("./il/tuo/path/bigscience_t0")
```
3. Ora quando sei offline, carica i tuoi file con [`PreTrainedModel.from_pretrained`] dalla directory specificata:
```py
>>> tokenizer = AutoTokenizer.from_pretrained("./il/tuo/path/bigscience_t0")
>>> model = AutoModel.from_pretrained("./il/tuo/path/bigscience_t0")
```
* Scarica in maniera programmatica i file con la libreria [huggingface_hub](https://github.com/huggingface/huggingface_hub/tree/main/src/huggingface_hub):
1. Installa la libreria `huggingface_hub` nel tuo ambiente virtuale:
```bash
python -m pip install huggingface_hub
```
2. Utilizza la funzione [`hf_hub_download`](https://huggingface.co/docs/hub/adding-a-library#download-files-from-the-hub) per scaricare un file in un path specifico. Per esempio, il seguente comando scarica il file `config.json` dal modello [T0](https://huggingface.co/bigscience/T0_3B) nel path che desideri:
```py
>>> from huggingface_hub import hf_hub_download
>>> hf_hub_download(repo_id="bigscience/T0_3B", filename="config.json", cache_dir="./il/tuo/path/bigscience_t0")
```
Una volta che il tuo file รจ scaricato e salvato in cache localmente, specifica il suo path locale per caricarlo e utilizzarlo:
```py
>>> from transformers import AutoConfig
>>> config = AutoConfig.from_pretrained("./il/tuo/path/bigscience_t0/config.json")
```
<Tip>
Fai riferimento alla sezione [How to download files from the Hub](https://huggingface.co/docs/hub/how-to-downstream) per avere maggiori dettagli su come scaricare modelli presenti sull Hub.
</Tip>
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/big_models.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Istanziare un big model
Quando vuoi utilizzare un modello preaddestrato (pretrained) molto grande, una sfida รจ minimizzare l'uso della RAM. Il workflow classico
in PyTorch รจ:
1. Crea il tuo modello con pesi casuali (random weights).
2. Carica i tuoi pesi preaddestrati.
3. Inserisci i pesi preaddestrati nel tuo modello casuale.
I passi 1 e 2 una versione completa del modello in memoria, in molti casi non รจ un problema, ma se il modello inizia a pesare diversi GigaBytes, queste due copie possono sturare la nostra RAM. Ancora peggio, se stai usando `torch.distributed` per seguire l'addestramento (training) in distribuito, ogni processo caricherร il modello preaddestrato e memorizzerร queste due copie nella RAM.
<Tip>
Nota che il modello creato casualmente รจ inizializzato con tensori "vuoti", che occupano spazio in memoria ma senza riempirlo (quindi i valori casuali sono quelli che si trovavano in questa porzione di memoria in un determinato momento). L'inizializzazione casuale che segue la distribuzione appropriata per il tipo di modello/parametri istanziato (come la distribuzione normale per le istanze) รจ eseguito solo dopo il passaggio 3 sui pesi non inizializzati, per essere piรน rapido possibile!
</Tip>
In questa guida, esploreremo le soluzioni che Transformers offre per affrontare questo problema. C'รจ da tenere in conto che questa รจ un'area in cui si sta attualmente sviluppando, quindi le API spiegate qui possono variare velocemente in futuro.
## Checkpoints condivisi
Dalla versione 4.18.0, i checkpoints dei modelli che occupano piรน di 10GB di spazio vengono automaticamente frammentati in piรน parti. Per quanto riguarda la possibilitร di avere un unico checkpoint quando si utilizza `model.save_pretrained(save_dir)`, si hanno diversi checkpoint parziali (ognuno con dimensione < 10GB) e un indice che mappa i nomi dei parametri ai file in cui sono memorizzati.
Puoi controllare la dimensione massima dopo la frammentazione con il parametro `max_shard_size`, nel prossimo esempio, useremo modelli di dimensioni normali con frammenti di piccoli dimensioni: prendiamo un modello BERT classico.
```py
from transformers import AutoModel
model = AutoModel.from_pretrained("bert-base-cased")
```
Se tu salvi usando [`~PreTrainedModel.save_pretrained`], avrai una nuova cartella con due file: il config del modello e i suoi pesi:
```py
>>> import os
>>> import tempfile
>>> with tempfile.TemporaryDirectory() as tmp_dir:
... model.save_pretrained(tmp_dir)
... print(sorted(os.listdir(tmp_dir)))
['config.json', 'pytorch_model.bin']
```
Adesso usiamo una dimensione massima di frammentazione di 200MB:
```py
>>> with tempfile.TemporaryDirectory() as tmp_dir:
... model.save_pretrained(tmp_dir, max_shard_size="200MB")
... print(sorted(os.listdir(tmp_dir)))
['config.json', 'pytorch_model-00001-of-00003.bin', 'pytorch_model-00002-of-00003.bin', 'pytorch_model-00003-of-00003.bin', 'pytorch_model.bin.index.json']
```
In aggiunta alla configurazione del modello, vediamo tre differenti file dei pesi, e un file `index.json` che รจ il nostro indice. Un checkpoint puรฒ essere ricaricato totalmente usando il metodo [`~PreTrainedModel.from_pretrained`]:
```py
>>> with tempfile.TemporaryDirectory() as tmp_dir:
... model.save_pretrained(tmp_dir, max_shard_size="200MB")
... new_model = AutoModel.from_pretrained(tmp_dir)
```
Il vantaggio principale di applicare questo metodo per modelli grandi รจ che durante il passo 2 del workflow illustrato in precedenza, ogni frammento del checkpoint viene caricato dopo il precedente, limitando l'utilizzo della RAM alla dimensione del modello piรน la dimensione del frammento piรน grande.
Dietro le quinte, il file indice รจ utilizzato per determinare quali chiavi sono nel checkpoint, e dove i corrispondenti pesi sono memorizzati. Possiamo caricare l'indice come un qualsiasi json e ottenere un dizionario:
```py
>>> import json
>>> with tempfile.TemporaryDirectory() as tmp_dir:
... model.save_pretrained(tmp_dir, max_shard_size="200MB")
... with open(os.path.join(tmp_dir, "pytorch_model.bin.index.json"), "r") as f:
... index = json.load(f)
>>> print(index.keys())
dict_keys(['metadata', 'weight_map'])
```
I metadati consistono solo nella dimensione totale del modello per ora. Abbiamo in programma di aggiungere altre informazioni in futuro:
```py
>>> index["metadata"]
{'total_size': 433245184}
```
La mappa dei pesi รจ la parte principale di questo indice, che mappa ogni nome dei parametri (si trova solitamente nei modelli PyTorch come `state_dict`) al file in cui รจ memorizzato:
```py
>>> index["weight_map"]
{'embeddings.LayerNorm.bias': 'pytorch_model-00001-of-00003.bin',
'embeddings.LayerNorm.weight': 'pytorch_model-00001-of-00003.bin',
...
```
Se vuoi caricare direttamente un checkpoint frammentato in un modello senza usare [`~PreTrainedModel.from_pretrained`] (come si farebbe con `model.load_state_dict()` per un checkpoint completo) devi usare [`~modeling_utils.load_sharded_checkpoint`]:
```py
>>> from transformers.modeling_utils import load_sharded_checkpoint
>>> with tempfile.TemporaryDirectory() as tmp_dir:
... model.save_pretrained(tmp_dir, max_shard_size="200MB")
... load_sharded_checkpoint(model, tmp_dir)
```
## Caricamento low memory
Frammentare i checkpoint l'utilizzo di memoria al passo 2 del workflow citato in precedenza, ma per utilizzare questo modello in un ambiente con poca memoria, consigliamo di utilizzare i nostri strumenti basati sulla libreria Accelerate.
Per ulteriori informazioni, leggere la seguente guida: [Large model loading using Accelerate](./main_classes/model#large-model-loading)
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/pipeline_tutorial.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Pipeline per l'inferenza
La [`pipeline`] rende semplice usare qualsiasi modello dal [Model Hub](https://huggingface.co/models) per fare inferenza su diversi compiti come generazione del testo, segmentazione di immagini e classificazione di audio. Anche se non hai esperienza con una modalitร specifica o non comprendi bene il codice che alimenta i modelli, รจ comunque possibile utilizzarli con l'opzione [`pipeline`]! Questa esercitazione ti insegnerร a:
* Usare una [`pipeline`] per fare inferenza.
* Usare uno specifico tokenizer o modello.
* Usare una [`pipeline`] per compiti che riguardano audio e video.
<Tip>
Dai un'occhiata alla documentazione di [`pipeline`] per una lista completa dei compiti supportati.
</Tip>
## Utilizzo della Pipeline
Nonostante ogni compito abbia una [`pipeline`] associata, รจ piรน semplice utilizzare l'astrazione generica della [`pipeline`] che contiene tutte quelle specifiche per ogni mansione. La [`pipeline`] carica automaticamente un modello predefinito e un tokenizer in grado di fare inferenza per il tuo compito.
1. Inizia creando una [`pipeline`] e specificando il compito su cui fare inferenza:
```py
>>> from transformers import pipeline
>>> generator = pipeline(task="text-generation")
```
2. Inserisci il testo in input nella [`pipeline`]:
```py
>>> generator(
... "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone"
... ) # doctest: +SKIP
[{'generated_text': 'Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone, Seven for the Iron-priests at the door to the east, and thirteen for the Lord Kings at the end of the mountain'}]
```
Se hai piรน di un input, inseriscilo in una lista:
```py
>>> generator(
... [
... "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone",
... "Nine for Mortal Men, doomed to die, One for the Dark Lord on his dark throne",
... ]
... ) # doctest: +SKIP
```
Qualsiasi parametro addizionale per il tuo compito puรฒ essere incluso nella [`pipeline`]. La mansione `text-generation` ha un metodo [`~generation.GenerationMixin.generate`] con diversi parametri per controllare l'output. Ad esempio, se desideri generare piรน di un output, utilizza il parametro `num_return_sequences`:
```py
>>> generator(
... "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone",
... num_return_sequences=2,
... ) # doctest: +SKIP
```
### Scegliere modello e tokenizer
La [`pipeline`] accetta qualsiasi modello dal [Model Hub](https://huggingface.co/models). Ci sono tag nel Model Hub che consentono di filtrare i modelli per attivitร . Una volta che avrai scelto il modello appropriato, caricalo usando la corrispondente classe `AutoModelFor` e [`AutoTokenizer`]. Ad esempio, carica la classe [`AutoModelForCausalLM`] per un compito di causal language modeling:
```py
>>> from transformers import AutoTokenizer, AutoModelForCausalLM
>>> tokenizer = AutoTokenizer.from_pretrained("distilgpt2")
>>> model = AutoModelForCausalLM.from_pretrained("distilgpt2")
```
Crea una [`pipeline`] per il tuo compito, specificando il modello e il tokenizer che hai caricato:
```py
>>> from transformers import pipeline
>>> generator = pipeline(task="text-generation", model=model, tokenizer=tokenizer)
```
Inserisci il testo di input nella [`pipeline`] per generare del testo:
```py
>>> generator(
... "Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone"
... ) # doctest: +SKIP
[{'generated_text': 'Three Rings for the Elven-kings under the sky, Seven for the Dwarf-lords in their halls of stone, Seven for the Dragon-lords (for them to rule in a world ruled by their rulers, and all who live within the realm'}]
```
## Audio pipeline
La flessibilitร della [`pipeline`] fa si che possa essere estesa ad attivitร sugli audio.
Per esempio, classifichiamo le emozioni in questo clip audio:
```py
>>> from datasets import load_dataset
>>> import torch
>>> torch.manual_seed(42) # doctest: +IGNORE_RESULT
>>> ds = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation")
>>> audio_file = ds[0]["audio"]["path"]
```
Trova un modello per la [classificazione audio](https://huggingface.co/models?pipeline_tag=audio-classification) sul Model Hub per eseguire un compito di riconoscimento automatico delle emozioni e caricalo nella [`pipeline`]:
```py
>>> from transformers import pipeline
>>> audio_classifier = pipeline(
... task="audio-classification", model="ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition"
... )
```
Inserisci il file audio nella [`pipeline`]:
```py
>>> preds = audio_classifier(audio_file)
>>> preds = [{"score": round(pred["score"], 4), "label": pred["label"]} for pred in preds]
>>> preds
[{'score': 0.1315, 'label': 'calm'}, {'score': 0.1307, 'label': 'neutral'}, {'score': 0.1274, 'label': 'sad'}, {'score': 0.1261, 'label': 'fearful'}, {'score': 0.1242, 'label': 'happy'}]
```
## Vision pipeline
Infine, usare la [`pipeline`] per le attivitร sulle immagini รจ praticamente la stessa cosa.
Specifica la tua attivitร e inserisci l'immagine nel classificatore. L'immagine puรฒ essere sia un link che un percorso sul tuo pc in locale. Per esempio, quale specie di gatto รจ raffigurata qui sotto?
```py
>>> from transformers import pipeline
>>> vision_classifier = pipeline(task="image-classification")
>>> preds = vision_classifier(
... images="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/pipeline-cat-chonk.jpeg"
... )
>>> preds = [{"score": round(pred["score"], 4), "label": pred["label"]} for pred in preds]
>>> preds
[{'score': 0.4335, 'label': 'lynx, catamount'}, {'score': 0.0348, 'label': 'cougar, puma, catamount, mountain lion, painter, panther, Felis concolor'}, {'score': 0.0324, 'label': 'snow leopard, ounce, Panthera uncia'}, {'score': 0.0239, 'label': 'Egyptian cat'}, {'score': 0.0229, 'label': 'tiger cat'}]
```
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/perf_train_cpu_many.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Addestramento effciente su multiple CPU
Quando l'addestramento su una singola CPU รจ troppo lento, possiamo usare CPU multiple. Quasta guida si concentra su DDP basato su PyTorch abilitando l'addetramento distribuito su CPU in maniera efficiente.
## Intelยฎ oneCCL Bindings per PyTorch
[Intelยฎ oneCCL](https://github.com/oneapi-src/oneCCL) (collective communications library) รจ una libreria per l'addestramento efficiente del deep learning in distribuito e implementa collettivi come allreduce, allgather, alltoall. Per maggiori informazioni su oneCCL, fai riferimento a [oneCCL documentation](https://spec.oneapi.com/versions/latest/elements/oneCCL/source/index.html) e [oneCCL specification](https://spec.oneapi.com/versions/latest/elements/oneCCL/source/index.html).
Il modulo `oneccl_bindings_for_pytorch` (`torch_ccl` precedentemente alla versione 1.12) implementa PyTorch C10D ProcessGroup API e puรฒ essere caricato dinamicamente com external ProcessGroup e funziona solo su piattaforma Linux al momento.
Qui trovi informazioni piรน dettagliate per [oneccl_bind_pt](https://github.com/intel/torch-ccl).
### Intelยฎ oneCCL Bindings per l'installazione PyTorch:
I file wheel sono disponibili per le seguenti versioni di Python:
| Extension Version | Python 3.6 | Python 3.7 | Python 3.8 | Python 3.9 | Python 3.10 |
| :---------------: | :--------: | :--------: | :--------: | :--------: | :---------: |
| 1.13.0 | | โ | โ | โ | โ |
| 1.12.100 | | โ | โ | โ | โ |
| 1.12.0 | | โ | โ | โ | โ |
| 1.11.0 | | โ | โ | โ | โ |
| 1.10.0 | โ | โ | โ | โ | |
```bash
pip install oneccl_bind_pt=={pytorch_version} -f https://developer.intel.com/ipex-whl-stable-cpu
```
dove `{pytorch_version}` deve essere la tua versione di PyTorch, per l'stanza 1.13.0.
Verifica altri approcci per [oneccl_bind_pt installation](https://github.com/intel/torch-ccl).
Le versioni di oneCCL e PyTorch devono combaciare.
<Tip warning={true}>
oneccl_bindings_for_pytorch 1.12.0 prebuilt wheel does not work with PyTorch 1.12.1 (it is for PyTorch 1.12.0)
PyTorch 1.12.1 should work with oneccl_bindings_for_pytorch 1.12.100
</Tip>
## Intelยฎ MPI library
Usa questa implementazione basata su standard MPI per fornire una architettura flessibile, efficiente, scalabile su cluster per Intelยฎ. Questo componente รจ parte di Intelยฎ oneAPI HPC Toolkit.
oneccl_bindings_for_pytorch รจ installato insieme al set di strumenti MPI. Necessitร di reperire l'ambiente prima di utilizzarlo.
per Intelยฎ oneCCL >= 1.12.0
```bash
oneccl_bindings_for_pytorch_path=$(python -c "from oneccl_bindings_for_pytorch import cwd; print(cwd)")
source $oneccl_bindings_for_pytorch_path/env/setvars.sh
```
per Intelยฎ oneCCL con versione < 1.12.0
```bash
torch_ccl_path=$(python -c "import torch; import torch_ccl; import os; print(os.path.abspath(os.path.dirname(torch_ccl.__file__)))")
source $torch_ccl_path/env/setvars.sh
```
#### Installazione IPEX:
IPEX fornisce ottimizzazioni delle prestazioni per l'addestramento della CPU sia con Float32 che con BFloat16; puoi fare riferimento a [single CPU section](./perf_train_cpu).
Il seguente "Utilizzo in Trainer" prende come esempio mpirun nella libreria Intelยฎ MPI.
## Utilizzo in Trainer
Per abilitare l'addestramento distribuito multi CPU nel Trainer con il ccl backend, gli utenti devono aggiungere **`--ddp_backend ccl`** negli argomenti del comando.
Vediamo un esempio per il [question-answering example](https://github.com/huggingface/transformers/tree/main/examples/pytorch/question-answering)
Il seguente comando abilita due processi sul nodo Xeon, con un processo in esecuzione per ogni socket. Le variabili OMP_NUM_THREADS/CCL_WORKER_COUNT possono essere impostate per una prestazione ottimale.
```shell script
export CCL_WORKER_COUNT=1
export MASTER_ADDR=127.0.0.1
mpirun -n 2 -genv OMP_NUM_THREADS=23 \
python3 run_qa.py \
--model_name_or_path bert-large-uncased \
--dataset_name squad \
--do_train \
--do_eval \
--per_device_train_batch_size 12 \
--learning_rate 3e-5 \
--num_train_epochs 2 \
--max_seq_length 384 \
--doc_stride 128 \
--output_dir /tmp/debug_squad/ \
--no_cuda \
--ddp_backend ccl \
--use_ipex
```
Il seguente comando abilita l'addestramento per un totale di quattro processi su due Xeon (node0 e node1, prendendo node0 come processo principale), ppn (processes per node) รจ impostato a 2, on un processo in esecuzione per ogni socket. Le variabili OMP_NUM_THREADS/CCL_WORKER_COUNT possono essere impostate per una prestazione ottimale.
In node0, รจ necessario creare un file di configurazione che contenga gli indirizzi IP di ciascun nodo (per esempio hostfile) e passare il percorso del file di configurazione come parametro.
```shell script
cat hostfile
xxx.xxx.xxx.xxx #node0 ip
xxx.xxx.xxx.xxx #node1 ip
```
A questo punto, esegui il seguente comando nel nodo0 e **4DDP** sarร abilitato in node0 e node1 con BF16 auto mixed precision:
```shell script
export CCL_WORKER_COUNT=1
export MASTER_ADDR=xxx.xxx.xxx.xxx #node0 ip
mpirun -f hostfile -n 4 -ppn 2 \
-genv OMP_NUM_THREADS=23 \
python3 run_qa.py \
--model_name_or_path bert-large-uncased \
--dataset_name squad \
--do_train \
--do_eval \
--per_device_train_batch_size 12 \
--learning_rate 3e-5 \
--num_train_epochs 2 \
--max_seq_length 384 \
--doc_stride 128 \
--output_dir /tmp/debug_squad/ \
--no_cuda \
--ddp_backend ccl \
--use_ipex \
--bf16
```
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/_toctree.yml
|
- sections:
- local: index
title: ๐ค Transformers
- local: quicktour
title: Tour rapido
- local: installation
title: Installazione
title: Iniziare
- sections:
- local: pipeline_tutorial
title: Pipeline per l'inferenza
- local: autoclass_tutorial
title: Carica istanze pre-allenate con AutoClass
- local: preprocessing
title: Preprocess
- local: training
title: Fine-tuning di un modello pre-addestrato
- local: accelerate
title: Allenamento distribuito con ๐ค Accelerate
- local: model_sharing
title: Condividere un modello
title: Esercitazione
- sections:
- local: create_a_model
title: Crea un'architettura personalizzata
- local: custom_models
title: Condividere modelli personalizzati
- local: run_scripts
title: Addestramento con script
- local: multilingual
title: Modelli multilingua per l'inferenza
- local: converting_tensorflow_models
title: Convertire modelli tensorflow
- local: serialization
title: Esporta modelli Transformers
- local: perf_train_cpu
title: Addestramento efficiente su CPU
- local: perf_train_cpu_many
title: Addestramento efficiente su multiple CPU
- local: perf_train_tpu
title: Addestramento su TPU
- local: perf_train_special
title: Addestramento su Hardware Specializzato
- local: perf_infer_cpu
title: Inferenza Efficiente su CPU
- local: perf_infer_gpu_one
title: Inferenza su una GPU
- local: perf_infer_gpu_many
title: Inferenza Efficiente su GPU Multiple
- local: perf_infer_special
title: Inferenza su Hardware Specializzato
- local: big_models
title: Istanziare un big model
- local: migration
title: Passaggio da pacchetti precedenti
- local: debugging
title: Debugging
title: Guide pratiche
- sections:
- local: add_new_pipeline
title: Come aggiungere una pipeline a ๐ค Transformers?
- local: add_new_model
title: Come aggiungere un modello a ๐ค Transformers?
- local: perf_hardware
title: Hardware ottimizzato per l'addestramento
- local: community
title: Risorse della comunitร
- local: pr_checks
title: Controlli su una Pull Request
title: Guide How-to
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/perf_train_cpu.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Addestramento efficiente su CPU
Questa guida si concentra su come addestrare in maniera efficiente grandi modelli su CPU.
## Mixed precision con IPEX
IPEX รจ ottimizzato per CPU con AVX-512 o superiore, e funziona per le CPU con solo AVX2. Pertanto, si prevede che le prestazioni saranno piรน vantaggiose per le le CPU Intel con AVX-512 o superiori, mentre le CPU con solo AVX2 (ad esempio, le CPU AMD o le CPU Intel piรน vecchie) potrebbero ottenere prestazioni migliori con IPEX, ma non sono garantite. IPEX offre ottimizzazioni delle prestazioni per l'addestramento della CPU sia con Float32 che con BFloat16. L'uso di BFloat16 รจ l'argomento principale delle seguenti sezioni.
Il tipo di dati a bassa precisione BFloat16 รจ stato supportato in modo nativo su 3rd Generation Xeonยฎ Scalable Processors (aka Cooper Lake) con AVX512 e sarร supportata dalla prossima generazione di Intelยฎ Xeonยฎ Scalable Processors con Intelยฎ Advanced Matrix Extensions (Intelยฎ AMX) instruction set con prestazioni ulteriormente migliorate. L'Auto Mixed Precision per il backende della CPU รจ stato abilitato da PyTorch-1.10. allo stesso tempo, il supporto di Auto Mixed Precision con BFloat16 per CPU e l'ottimizzazione degli operatori BFloat16 รจ stata abilitata in modo massiccio in Intelยฎ Extension per PyTorch, and parzialmente aggiornato al branch master di PyTorch. Gli utenti possono ottenere prestazioni migliori ed users experience con IPEX Auto Mixed Precision..
Vedi informazioni piรน dettagliate su [Auto Mixed Precision](https://intel.github.io/intel-extension-for-pytorch/cpu/latest/tutorials/features/amp.html).
### Installazione di IPEX:
Il rilascio di IPEX segue quello di PyTorch, da installare via pip:
| PyTorch Version | IPEX version |
| :---------------: | :----------: |
| 1.13 | 1.13.0+cpu |
| 1.12 | 1.12.300+cpu |
| 1.11 | 1.11.200+cpu |
| 1.10 | 1.10.100+cpu |
```bash
pip install intel_extension_for_pytorch==<version_name> -f https://developer.intel.com/ipex-whl-stable-cpu
```
Vedi altri approcci per [IPEX installation](https://intel.github.io/intel-extension-for-pytorch/cpu/latest/tutorials/installation.html).
### Utilizzo nel Trainer
Per abilitare la auto mixed precision con IPEX in Trainer, l'utende dovrebbe aggiungere `use_ipex`, `bf16` e `no_cuda` negli argomenti del comando di addestramento.
Vedi un sempio di un caso d'uso [Transformers question-answering](https://github.com/huggingface/transformers/tree/main/examples/pytorch/question-answering)
- Training with IPEX using BF16 auto mixed precision on CPU:
<pre> python run_qa.py \
--model_name_or_path bert-base-uncased \
--dataset_name squad \
--do_train \
--do_eval \
--per_device_train_batch_size 12 \
--learning_rate 3e-5 \
--num_train_epochs 2 \
--max_seq_length 384 \
--doc_stride 128 \
--output_dir /tmp/debug_squad/ \
<b>--use_ipex \</b>
<b>--bf16 --no_cuda</b></pre>
### Esempi pratici
Blog: [Accelerating PyTorch Transformers with Intel Sapphire Rapids](https://huggingface.co/blog/intel-sapphire-rapids)
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/migration.md
|
<!---
Copyright 2020 The HuggingFace Team. Tutti i diritti riservati.
Concesso in licenza in base alla Licenza Apache, Versione 2.0 (la "Licenza");
non รจ possibile utilizzare questo file se non in conformitร con la Licenza.
ร possibile ottenere una copia della Licenza all'indirizzo
http://www.apache.org/licenses/LICENSE-2.0
A meno che non sia richiesto dalla legge applicabile o concordato per iscritto, il software
distribuito con la Licenza รจ distribuito su BASE "COSร COM'ร",
SENZA GARANZIE O CONDIZIONI DI ALCUN TIPO, espresse o implicite.
Per la lingua specifica vedi la Licenza che regola le autorizzazioni e
le limitazioni ai sensi della STESSA.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Migrazione da pacchetti precedenti
## Migrazione da transformers `v3.x` a `v4.x`
Un paio di modifiche sono state introdotte nel passaggio dalla versione 3 alla versione 4. Di seguito รจ riportato un riepilogo delle
modifiche previste:
#### 1. AutoTokenizer e pipeline ora utilizzano tokenizer veloci (rust) per impostazione predefinita.
I tokenizer python e rust hanno all'incirca le stesse API, ma i tokenizer rust hanno un set di funzionalitร piรน completo.
Ciรฒ introduce due modifiche sostanziali:
- La gestione dei token in overflow tra i tokenizer Python e Rust รจ diversa.
- I tokenizers di rust non accettano numeri interi nei metodi di codifica.
##### Come ottenere lo stesso comportamento di v3.x in v4.x
- Le pipeline ora contengono funzionalitร aggiuntive pronte all'uso. Vedi la [pipeline di classificazione dei token con il flag `grouped_entities`](main_classes/pipelines#transformers.TokenClassificationPipeline).
- Gli auto-tokenizer ora restituiscono tokenizer rust. Per ottenere invece i tokenizer python, l'utente deve usare il flag `use_fast` impostandolo `False`:
Nella versione `v3.x`:
```py
from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("bert-base-cased")
```
per ottenere lo stesso nella versione `v4.x`:
```py
from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained("bert-base-cased", use_fast=False)
```
#### 2. SentencePiece รจ stato rimosso dalle dipendenze richieste
Il requisito sulla dipendenza SentencePiece รจ stato rimosso da `setup.py`. ร stato fatto per avere un canale su anaconda cloud senza basarsi su `conda-forge`. Ciรฒ significa che i tokenizer che dipendono dalla libreria SentencePiece non saranno disponibili con un'installazione standard di `transformers`.
Ciรฒ include le versioni **lente** di:
- `XLNetTokenizer`
- `AlbertTokenizer`
- `CamembertTokenizer`
- `MBartTokenizer`
- `PegasusTokenizer`
- `T5Tokenizer`
- `ReformerTokenizer`
- `XLMRobertaTokenizer`
##### Come ottenere lo stesso comportamento della v3.x nella v4.x
Per ottenere lo stesso comportamento della versione `v3.x`, devi installare anche `sentencepiece`:
Nella versione `v3.x`:
```bash
pip install transformers
```
per ottenere lo stesso nella versione `v4.x`:
```bash
pip install transformers[sentencepiece]
```
o
```bash
pip install transformers stentencepiece
```
#### 3. L'architettura delle repo รจ stato aggiornata in modo che ogni modello abbia la propria cartella
Con lโaggiunta di nuovi modelli, il numero di file nella cartella `src/transformers` continua a crescere e diventa piรน difficile navigare e capire. Abbiamo fatto la scelta di inserire ogni modello e i file che lo accompagnano nelle proprie sottocartelle.
Si tratta di una modifica sostanziale in quanto l'importazione di layer intermedi utilizzando direttamente il modulo di un modello deve essere eseguita tramite un percorso diverso.
##### Come ottenere lo stesso comportamento della v3.x nella v4.x
Per ottenere lo stesso comportamento della versione `v3.x`, devi aggiornare il percorso utilizzato per accedere ai layer.
Nella versione `v3.x`:
```bash
from transformers.modeling_bert import BertLayer
```
per ottenere lo stesso nella versione `v4.x`:
```bash
from transformers.models.bert.modeling_bert import BertLayer
```
#### 4. Impostare l'argomento `return_dict` su `True` per impostazione predefinita
L'[argomento `return_dict`](main_classes/output) abilita la restituzione di oggetti python dict-like contenenti gli output del modello, invece delle tuple standard. Questo oggetto รจ self-documented poichรฉ le chiavi possono essere utilizzate per recuperare valori, comportandosi anche come una tupla e gli utenti possono recuperare oggetti per indexing o slicing.
Questa รจ una modifica sostanziale poichรฉ la tupla non puรฒ essere decompressa: `value0, value1 = outputs` non funzionerร .
##### Come ottenere lo stesso comportamento della v3.x nella v4.x
Per ottenere lo stesso comportamento della versione `v3.x`, specifica l'argomento `return_dict` come `False`, sia nella configurazione del modello che nel passaggio successivo.
Nella versione `v3.x`:
```bash
model = BertModel.from_pretrained("bert-base-cased")
outputs = model(**inputs)
```
per ottenere lo stesso nella versione `v4.x`:
```bash
model = BertModel.from_pretrained("bert-base-cased")
outputs = model(**inputs, return_dict=False)
```
o
```bash
model = BertModel.from_pretrained("bert-base-cased", return_dict=False)
outputs = model(**inputs)
```
#### 5. Rimozione di alcuni attributi deprecati
Gli attributi sono stati rimossi se deprecati da almeno un mese. L'elenco completo degli attributi obsoleti รจ disponibile in [#8604](https://github.com/huggingface/transformers/pull/8604).
Ecco un elenco di questi attributi/metodi/argomenti e quali dovrebbero essere le loro sostituzioni:
In diversi modelli, le etichette diventano coerenti con gli altri modelli:
- `masked_lm_labels` diventa `labels` in `AlbertForMaskedLM` e `AlbertForPreTraining`.
- `masked_lm_labels` diventa `labels` in `BertForMaskedLM` e `BertForPreTraining`.
- `masked_lm_labels` diventa `labels` in `DistilBertForMaskedLM`.
- `masked_lm_labels` diventa `labels` in `ElectraForMaskedLM`.
- `masked_lm_labels` diventa `labels` in `LongformerForMaskedLM`.
- `masked_lm_labels` diventa `labels` in `MobileBertForMaskedLM`.
- `masked_lm_labels` diventa `labels` in `RobertaForMaskedLM`.
- `lm_labels` diventa `labels` in `BartForConditionalGeneration`.
- `lm_labels` diventa `labels` in `GPT2DoubleHeadsModel`.
- `lm_labels` diventa `labels` in `OpenAIGPTDoubleHeadsModel`.
- `lm_labels` diventa `labels` in `T5ForConditionalGeneration`.
In diversi modelli, il meccanismo di memorizzazione nella cache diventa coerente con gli altri:
- `decoder_cached_states` diventa `past_key_values` in tutti i modelli BART-like, FSMT e T5.
- `decoder_past_key_values` diventa `past_key_values` in tutti i modelli BART-like, FSMT e T5.
- `past` diventa `past_key_values` in tutti i modelli CTRL.
- `past` diventa `past_key_values` in tutti i modelli GPT-2.
Per quanto riguarda le classi tokenizer:
- L'attributo tokenizer `max_len` diventa `model_max_length`.
- L'attributo tokenizer `return_lengths` diventa `return_length`.
- L'argomento di codifica del tokenizer `is_pretokenized` diventa `is_split_into_words`.
Per quanto riguarda la classe `Trainer`:
- L'argomento `tb_writer` di `Trainer` รจ stato rimosso in favore della funzione richiamabile `TensorBoardCallback(tb_writer=...)`.
- L'argomento `prediction_loss_only` di `Trainer` รจ stato rimosso in favore dell'argomento di classe `args.prediction_loss_only`.
- L'attributo `data_collator` di `Trainer` sarร richiamabile.
- Il metodo `_log` di `Trainer` รจ deprecato a favore di `log`.
- Il metodo `_training_step` di `Trainer` รจ deprecato a favore di `training_step`.
- Il metodo `_prediction_loop` di `Trainer` รจ deprecato a favore di `prediction_loop`.
- Il metodo `is_local_master` di `Trainer` รจ deprecato a favore di `is_local_process_zero`.
- Il metodo `is_world_master` di `Trainer` รจ deprecato a favore di `is_world_process_zero`.
Per quanto riguarda la classe `TFTrainer`:
- L'argomento `prediction_loss_only` di `TFTrainer` รจ stato rimosso a favore dell'argomento di classe `args.prediction_loss_only`.
- Il metodo `_log` di `Trainer` รจ deprecato a favore di `log`.
- Il metodo `_prediction_loop` di `TFTrainer` รจ deprecato a favore di `prediction_loop`.
- Il metodo `_setup_wandb` di `TFTrainer` รจ deprecato a favore di `setup_wandb`.
- Il metodo `_run_model` di `TFTrainer` รจ deprecato a favore di `run_model`.
Per quanto riguarda la classe `TrainingArguments`:
- L'argomento `evaluate_during_training` di `TrainingArguments` รจ deprecato a favore di `evaluation_strategy`.
Per quanto riguarda il modello Transfo-XL:
- L'attributo di configurazione `tie_weight` di Transfo-XL diventa `tie_words_embeddings`.
- Il metodo di modellazione `reset_length` di Transfo-XL diventa `reset_memory_length`.
Per quanto riguarda le pipeline:
- L'argomento `topk` di `FillMaskPipeline` diventa `top_k`.
## Passaggio da pytorch-transformers a ๐ค Transformers
Ecco un breve riepilogo di ciรฒ a cui prestare attenzione durante il passaggio da `pytorch-transformers` a ๐ค Transformers.
### Lโordine posizionale di alcune parole chiave di input dei modelli (`attention_mask`, `token_type_ids`...) รจ cambiato
Per usare Torchscript (vedi #1010, #1204 e #1195) l'ordine specifico delle **parole chiave di input** di alcuni modelli (`attention_mask`, `token_type_ids`...) รจ stato modificato.
Se inizializzavi i modelli usando parole chiave per gli argomenti, ad esempio `model(inputs_ids, attention_mask=attention_mask, token_type_ids=token_type_ids)`, questo non dovrebbe causare alcun cambiamento.
Se inizializzavi i modelli con input posizionali per gli argomenti, ad esempio `model(inputs_ids, attention_mask, token_type_ids)`, potrebbe essere necessario ricontrollare l'ordine esatto degli argomenti di input.
## Migrazione da pytorch-pretrained-bert
Ecco un breve riepilogo di ciรฒ a cui prestare attenzione durante la migrazione da `pytorch-pretrained-bert` a ๐ค Transformers
### I modelli restituiscono sempre `tuple`
La principale modifica di rilievo durante la migrazione da `pytorch-pretrained-bert` a ๐ค Transformers รจ che il metodo dei modelli di previsione dร sempre una `tupla` con vari elementi a seconda del modello e dei parametri di configurazione.
Il contenuto esatto delle tuple per ciascun modello รจ mostrato in dettaglio nelle docstring dei modelli e nella [documentazione](https://huggingface.co/transformers/).
In quasi tutti i casi, andrร bene prendendo il primo elemento dell'output come quello che avresti precedentemente utilizzato in `pytorch-pretrained-bert`.
Ecco un esempio di conversione da `pytorch-pretrained-bert`
a ๐ค Transformers per un modello di classificazione `BertForSequenceClassification`:
```python
# Carichiamo il nostro modello
model = BertForSequenceClassification.from_pretrained("bert-base-uncased")
# Se usavi questa riga in pytorch-pretrained-bert :
loss = model(input_ids, labels=labels)
# Ora usa questa riga in ๐ค Transformers per estrarre la perdita dalla tupla di output:
outputs = model(input_ids, labels=labels)
loss = outputs[0]
# In ๐ค Transformers puoi anche avere accesso ai logit:
loss, logits = outputs[:2]
# Ed anche agli attention weight se configuri il modello per restituirli (e anche altri output, vedi le docstring e la documentazione)
model = BertForSequenceClassification.from_pretrained(" bert-base-uncased", output_attentions=True)
outputs = model(input_ids, labels=labels)
loss, logits, attentions = outputs
```
### Serializzazione
Modifica sostanziale nel metodo `from_pretrained()`:
1. I modelli sono ora impostati in modalitร di valutazione in maniera predefinita quando usi il metodo `from_pretrained()`. Per addestrarli non dimenticare di riportarli in modalitร di addestramento (`model.train()`) per attivare i moduli di dropout.
2. Gli argomenti aggiuntivi `*inputs` e `**kwargs` forniti al metodo `from_pretrained()` venivano passati direttamente al metodo `__init__()` della classe sottostante del modello. Ora sono usati per aggiornare prima l'attributo di configurazione del modello, che puรฒ non funzionare con le classi del modello derivate costruite basandosi sui precedenti esempi di `BertForSequenceClassification`. Piรน precisamente, gli argomenti posizionali `*inputs` forniti a `from_pretrained()` vengono inoltrati direttamente al metodo `__init__()` del modello mentre gli argomenti keyword `**kwargs` (i) che corrispondono agli attributi della classe di configurazione, vengono utilizzati per aggiornare tali attributi (ii) che non corrispondono ad alcun attributo della classe di configurazione, vengono inoltrati al metodo `__init__()`.
Inoltre, sebbene non si tratti di una modifica sostanziale, i metodi di serializzazione sono stati standardizzati e probabilmente dovresti passare al nuovo metodo `save_pretrained(save_directory)` se prima usavi qualsiasi altro metodo di serializzazione.
Ecco un esempio:
```python
### Carichiamo un modello e un tokenizer
model = BertForSequenceClassification.from_pretrained("bert-base-uncased")
tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")
### Facciamo fare alcune cose al nostro modello e tokenizer
# Es: aggiungiamo nuovi token al vocabolario e agli embending del nostro modello
tokenizer.add_tokens(["[SPECIAL_TOKEN_1]", "[SPECIAL_TOKEN_2]"])
model.resize_token_embeddings(len(tokenizer))
# Alleniamo il nostro modello
train(model)
### Ora salviamo il nostro modello e il tokenizer in una cartella
model.save_pretrained("./my_saved_model_directory/")
tokenizer.save_pretrained("./my_saved_model_directory/")
### Ricarichiamo il modello e il tokenizer
model = BertForSequenceClassification.from_pretrained("./my_saved_model_directory/")
tokenizer = BertTokenizer.from_pretrained("./my_saved_model_directory/")
```
### Ottimizzatori: BertAdam e OpenAIAdam ora sono AdamW, lo scheduling รจ quello standard PyTorch
I due ottimizzatori precedenti inclusi, `BertAdam` e `OpenAIAdam`, sono stati sostituiti da un singolo `AdamW` che presenta alcune differenze:
- implementa solo la correzione del weights decay,
- lo scheduling ora รจ esterno (vedi sotto),
- anche il gradient clipping ora รจ esterno (vedi sotto).
Il nuovo ottimizzatore `AdamW` corrisponde alle API di `Adam` di PyTorch e ti consente di utilizzare metodi PyTorch o apex per lo scheduling e il clipping.
Lo scheduling รจ ora standard [PyTorch learning rate schedulers](https://pytorch.org/docs/stable/optim.html#how-to-adjust-learning-rate) e non fanno piรน parte dell'ottimizzatore.
Ecco un esempio di linear warmup e decay con `BertAdam` e con `AdamW`:
```python
# Parametri:
lr = 1e-3
max_grad_norm = 1.0
num_training_steps = 1000
num_warmup_steps = 100
warmup_proportion = float( num_warmup_steps) / float(num_training_steps) # 0.1
### In precedenza l'ottimizzatore BertAdam veniva istanziato in questo modo:
optimizer = BertAdam(
model.parameters(),
lr=lr,
schedule="warmup_linear",
warmup=warmup_proportion,
num_training_steps=num_training_steps,
)
### e usato in questo modo:
for batch in train_data:
loss = model(batch)
loss.backward()
optimizer.step()
### In ๐ค Transformers, ottimizzatore e schedule sono divisi e usati in questo modo:
optimizer = AdamW(
model.parameters(), lr=lr, correct_bias=False
) # Per riprodurre il comportamento specifico di BertAdam impostare correct_bias=False
scheduler = get_linear_schedule_with_warmup(
optimizer, num_warmup_steps=num_warmup_steps, num_training_steps=num_training_steps
) # PyTorch scheduler
### e va usato cosรฌ:
for batch in train_data:
loss = model(batch)
loss.backward()
torch.nn.utils.clip_grad_norm_(
model.parameters(), max_grad_norm
) # Gradient clipping non รจ piรน in AdamW (quindi puoi usare amp senza problemi)
optimizer.step()
scheduler.step()
```
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/add_new_model.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Come aggiungere un modello a ๐ค Transformers?
Aggiungere un nuovo modello รฉ spesso difficile e richiede una profonda conoscenza della libreria ๐ค Transformers e anche
della repository originale del modello. A Hugging Face cerchiamo di dare alla community sempre piรบ poteri per aggiungere
modelli independentemente. Quindi, per alcuni nuovi modelli che la community vuole aggiungere a ๐ค Transformers, abbiamo
creato una specifica *call-for-model-addition* che spiega passo dopo passo come aggiungere il modello richiesto. Con
questo *call-for-model-addition* vogliamo insegnare a volenterosi e esperti collaboratori della community come implementare
un modello in ๐ค Transformers.
Se questo รฉ qualcosa che puรฒ interessarvi, siete liberi di controllare l'attuale โcalls-for-model-additionโ [qui](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model/open_model_proposals/README.md)
e contattarci.
Se il modello sarร selezionato, allora potrete lavorare insieme a un membro di Hugging Face per integrare il modello in ๐ค
Transformers. Cosรฌ facendo, ci guadagnerai in una comprensione totale, sia teorica che pratica, del modello proposto. Inoltre,
sarai l'artefice di un importante contributo open-source a ๐ค Transformers. Durante l'implementazione avrai l'opportunitร di:
- ottenere piรน comprensione delle best practices in open-source
- capire i principi di design di una della librerie NLP piรน popolari
- capire come efficientemente testare complessi modelli NLP
- capire come integrare utilit Python come `black`, `ruff`, `make fix-copies` in una libreria per garantire sempre di avere un codice leggibile e pulito
Siamo anche contenti se vuoi aggiungere un modello che non puรฒ essere trovato nella cartella โcalls-for-model-additionโ.
Le seguenti sezioni spiegano in dettaglio come aggiungere un nuovo modello. Puรฒ anche essere molto utile controllare modelli
giร aggiunti [qui](https://github.com/huggingface/transformers/pulls?q=is%3Apr+label%3A%22PR+for+Model+Addition%22+is%3Aclosed),
per capire se richiamano il modello che vorreste aggiungere.
Per cominciare, vediamo una panoramica general della libreria Transformers.
## Panoramica generale su ๐ค Transformers
Prima di tutto, vediamo in generale ๐ค Transformers. ๐ค Transformers รฉ una libreria molto strutturata, quindi
puร essere che a volte ci sia un disaccordo con alcune filosofie della libreria o scelte di design. Dalla nostra esperienza,
tuttavia, abbiamo trovato che le scelte fondamentali di design della libreria sono cruciali per usare ๐ค Transformers efficacemente
su larga scala, mantenendo i costi a un livello accettabile.
Un buon primo punto di partenza per capire al meglio la libreria รฉ leggere la [documentazione sulla nostra filosofia](filosofia)
Da qui, ci sono alcune scelte sul modo di lavorare che cerchiamo di applicare a tutti i modelli:
- La composizione รฉ generalmente favorita sulla sovra-astrazione
- Duplicare il codice non รฉ sempre male, soprattutto se migliora notevolmente la leggibilitร e accessibilitร del modello
- Tutti i files creati per il nuovo modello devono il piu possibile "compatti". Questo vuol dire che quando qualcuno leggerรก il codice
di uno specifico modello, potrรก vedere solo il corrispettivo file `modeling_....py` senza avere multiple dipendenze.
La cosa piรบ importante, รฉ che consideriamo la libreria non solo un mezzo per dare un prodotto, *per esempio* dare la possibilitร
di usare BERT per inferenza, ma รฉ anche il prodotto reale che noi vogliamo migliorare sempre piรน. Quindi, quando aggiungi
un modello, non sei solo la persona che userร il modello, ma rappresenti anche tutti coloro che leggeranno,
cercheranno di capire e modificare il tuo modello.
Tenendo questi principi in mente, immergiamoci nel design generale della libreria.
### Panoramica sui modelli
Per aggiungere con successo un modello, รฉ importante capire l'interazione tra il tuo modello e la sua configurazione,
[`PreTrainedModel`], e [`PretrainedConfig`]. Per dare un esempio, chiameremo il modello da aggiungere a ๐ค Transformers
`BrandNewBert`.
Diamo un'occhiata:
<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers_overview.png"/>
Come potete vedere, ci basiamo sull'ereditarietร in ๐ค Transformers, tenendo perรฒ il livello di astrazione a un minimo
assoluto. Non ci sono mai piรน di due livelli di astrazione per ogni modello nella libreria. `BrandNewBertModel` eredita
da `BrandNewBertPreTrainedModel` che, a sua volta, eredita da [`PreTrainedModel`] - semplice no?
Come regola generale, vogliamo essere sicuri che un nuovo modello dipenda solo da [`PreTrainedModel`]. Le funzionalitร
importanti che sono automaticamente conferite a ogni nuovo modello sono [`~PreTrainedModel.from_pretrained`]
e [`~PreTrainedModel.save_pretrained`], che sono usate per serializzazione e deserializzazione. Tutte le altre importanti
funzionalitร , come ad esempio `BrandNewBertModel.forward` devono essere definite completamente nel nuovo script
`modeling_brand_new_bert.py`. Inoltre, vogliamo essere sicuri che un modello con uno specifico head layer, come
`BrandNewBertForMaskedLM` non erediti da `BrandNewBertModel`, ma piuttosto usi `BrandNewBertModel`
come componente che puรฒ essere chiamata nel passaggio forward per mantenere il livello di astrazione basso. Ogni
nuovo modello richieste una classe di configurazione, chiamata `BrandNewBertConfig`. Questa configurazione รฉ sempre
mantenuta come un attributo in [`PreTrainedModel`], e quindi puรฒ essere accessibile tramite l'attributo `config`
per tutte le classi che ereditano da `BrandNewBertPreTrainedModel`:
```python
model = BrandNewBertModel.from_pretrained("brandy/brand_new_bert")
model.config # il modello ha accesso al suo config
```
Analogamente al modello, la configurazione eredita le funzionalitร base di serializzazione e deserializzazione da
[`PretrainedConfig`]. ร da notare che la configurazione e il modello sono sempre serializzati in due formati differenti -
il modello รฉ serializzato in un file *pytorch_model.bin* mentre la configurazione con *config.json*. Chiamando
[`~PreTrainedModel.save_pretrained`] automaticamente chiamerร [`~PretrainedConfig.save_pretrained`], cosicchรฉ sia il
modello che la configurazione siano salvati.
### Stile per il codice
Quando codifichi un nuovo modello, tieni presente che Transformers ha una sua struttura di fondo come libreria, perciรฒ
ci sono alcuni fatti da considerare su come scrivere un codice :-)
1. Il forward pass del tuo modello dev'essere scritto completamente nel file del modello, mentre dev'essere indipendente
da altri modelli nella libreria. Se vuoi riutilizzare un blocco di codice da un altro modello, copia e incolla il codice con un commento `# Copied from` in cima al codice (guarda [qui](https://github.com/huggingface/transformers/blob/v4.17.0/src/transformers/models/roberta/modeling_roberta.py#L160)
per un ottimo esempio).
2. Il codice dev'essere interamente comprensibile, anche da persone che non parlano in inglese. Questo significa che le
variabili devono avere un nome descrittivo e bisogna evitare abbreviazioni. Per esempio, `activation` รฉ molto meglio
che `act`. Le variabili con una lettera sono da evitare fortemente, almeno che non sia per un indce in un for loop.
3. Generamente รฉ meglio avere un codice esplicito e piรบ lungo che un codice corto e magico.
4. Evita di subclassare `nn.Sequential` in Pytorch, puoi subclassare `nn.Module` e scrivere il forward pass, cosicchรฉ
chiunque puรฒ effettuare debug sul tuo codice, aggiungendo print o breaking points.
5. La tua function-signature dev'essere type-annoted. Per il resto, รฉ meglio preferire variabili con un nome accettabile
piuttosto che annotazioni per aumentare la comprensione e leggibilitร del codice.
### Panoramica sui tokenizers
Questa sezione sarร creata al piu presto :-(
## Aggiungere un modello a ๐ค Transformers passo dopo passo
Ci sono differenti modi per aggiungere un modello a Hugging Face. Qui trovi una lista di blog posts da parte della community su come aggiungere un modello:
1. [Aggiungere GPT2](https://medium.com/huggingface/from-tensorflow-to-pytorch-265f40ef2a28) scritto da [Thomas](https://huggingface.co/thomwolf)
2. [Aggiungere WMT19 MT](https://huggingface.co/blog/porting-fsmt) scritto da [Stas](https://huggingface.co/stas)
Per esperienza, possiamo dirti che quando si aggiunge un modello รฉ meglio tenere a mente le seguenti considerazioni:
- Non sfondare una porta giรก aperta! La maggior parte del codice che aggiungerai per un nuovo modello ๐ค Transformers
esiste giร da qualche parte in ๐ค Transformers. Prendi un po' di tempo per trovare codici simili in modelli e tokenizers esistenti e fare un copia-incolla. Ricorda che [grep](https://www.gnu.org/software/grep/) e [rg](https://github.com/BurntSushi/ripgrep) sono tuoi buoni amici. Inoltre, ricorda che puรณ essere molto probabile che il tokenizer per il tuo modello sia basato sull'implementazione di un altro modello, e il codice del tuo modello stesso su un altro ancora. *Per esempio* il modello FSMT รฉ basato su BART, mentre il tokenizer di FSMT รฉ basato su XLM.
- Ricorda che qui รฉ piu una sfida ingegneristica che scientifica. Spendi piรบ tempo per create un efficiente ambiente di debugging piuttosto che cercare di capire tutti gli aspetti teorici dell'articolo del modello.
- Chiedi aiuto se sei in panne! I modelli sono la parte principale di ๐ค Transformers, perciรฒ qui a Hugging Face siamo piรน che contenti di aiutarti in ogni passo per aggiungere il tuo modello. Non esitare a chiedere se vedi che non riesci a progredire.
Di seguito, diamo una ricetta generale per aiutare a portare un modello in ๐ค Transformers.
La lista seguente รฉ un sommario di tutto quello che รฉ stato fatto per aggiungere un modello, e puรฒ essere usata come To-Do List:
- 1. โ (Opzionale) Capire gli aspetti teorici del modello
- 2. โ Preparare l'ambiente dev per transformers
- 3. โ Preparare l'ambiente debugging della repository originale
- 4. โ Create uno script che gestisca con successo il forward pass usando la repository originale e checkpoint
- 5. โ Aggiungere con successo lo scheletro del modello a Transformers
- 6. โ Convertire i checkpoint original a Transformers checkpoint
- 7. โ Effettuare con successo la forward pass in Transformers, di modo che dia un output identico al checkpoint originale
- 8. โ Finire i tests per il modello in Transformers
- 9. โ Aggiungere con successo Tokenizer in Transformers
- 10. โ Testare e provare gli integration tests da capo a fine
- 11. โ Completare i docs
- 12. โ Caricare i moedl weights all'hub
- 13. โ Sottomettere una pull request
- 14. โ (Opzionale) Aggiungere un notebook con una demo
Per cominciare di solito consigliamo `BrandNewBert`, partendo dalla teoria, di modo da avere una buona comprensione della teoria generale. TUttavia, se preferisci imparare l'aspetto teorico del modello mentre *lavori* sul modello รฉ ok immergersi direttamente nel codice di `BrandNewBert`. Questa opzione puรณ essere buona se le tue skills ingegneristiche sono meglio che quelle teoriche, o se il paper `BrandNewBert` ti dรก problemi, o se semplicemente ti piace programmare piรบ che leggere articoli scientifici.
### 1. (Opzionale) Aspetti teorici di BrandNewBert
Allora con calma, prendi un po' di tempo per leggere l'articolo su *BrandNewBert* . Sicuramente, alcune sezioni dell'articolo sono molto complesse, ma non preoccuparti! L'obiettivo non รฉ avere una compresione immensa della teoria alla base, ma estrarre le informazioni necessarie per re-implementare con successo il modello in ๐ค Transformers. Quindi, non impazzire sugli aspetti teorici, ma piuttosto focalizzati su quelli pratici, ossia:
- Che tipo di modello รฉ *brand_new_bert*? ร solo un encoder in stile BERT? O tipo decoder come GPT2? O encoder e decoder stile BART? Dai un'occhiata a [model_summary](model_summary) se non sei famigliare con le differenze tra questi modelli
- Quali sono le applicazioni di *brand_new_bert*? Classificazione di testo? Generazione di testo? O per tasks del genere seq2seq?
- Quali sono le nuove aggiunte al modello che lo rendono diverso da BERT/GPT-2/BART?
- Quali modelli estistenti in [๐ค Transformers models](https://huggingface.co/transformers/#contents) sono molto simili a *brand_new_bert*?
- Che tipo di tokenizer si usa in questo caso? Un sentencepiece tokenizer? O un word piece tokenizer? Il tokenizer รฉ lo stesso di BERT o BART?
Una volta che senti che hai avuto una bella overview dell'architettura del modello, puoi scrivere senza problemi al team di Hugging Face per ogni domanda che tu hai. Questo puรณ includere domande sull'architettura del modello, o sull'attention layer, etc. Saremo molto felici di aiutarti :)
### 2. Prepare il tuo ambiente
1. Forka la [repository](https://github.com/huggingface/transformers) cliccando sul tasto โFork' nella pagina della repository. Questo crea una copia del codice nel tuo account GitHub
2. Clona il tuo fork `transfomers` sul tuo dico locale, e aggiungi la repository base come remota:
```bash
git clone https://github.com/[your Github handle]/transformers.git
cd transformers
git remote add upstream https://github.com/huggingface/transformers.git
```
3. Crea un ambiente di sviluppo, per esempio tramite questo comando:
```bash
python -m venv .env
source .env/bin/activate
pip install -e ".[dev]"
```
quindi torna alla directory principale:
```bash
cd ..
```
4. Attenzione, raccomandiamo di aggiungere la versione di PyTorch di *brand_new_bert* a Transfomers. Per installare PyTorch, basta seguire queste istruzioni https://pytorch.org/get-started/locally/.
**Nota bene:** Non c'รฉ bisogno di installare o avere installato CUDA. Il nuovo modello puรฒ funzionare senza problemi su una CPU.
5. Per trasferire *brand_new_bert* To port *brand_new_bert* avrai bisogno anche accesso alla sua repository originale:
```bash
git clone https://github.com/org_that_created_brand_new_bert_org/brand_new_bert.git
cd brand_new_bert
pip install -e .
```
Ok, ora hai un ambiente di sviluppo per portare *brand_new_bert* in ๐ค Transformers.
### 3.-4. Provare un pretrained checkpoint usando la repo originale
Per cominciare, comincerai a lavorare sulla repo originale di *brand_new_bert*. Come spesso accade, l'implementazione originale รฉ molto sullo stile "ricerca". Questo significa che a volte la documentazione non รฉ al top, magari manca qualche cosa e il codice puรณ essere difficile da capire. Tuttavia, questa รฉ e dev'essere la motivazione per reimplementare *brand_new_bert*. In Hugging Face, uno degli obiettivi principali รฉ di *mettere le persone sulle spalle dei giganti*, il che si traduce, in questo contesto, di prendere un modello funzionante e riscriverlo e renderlo il piรบ possibile **accessibile, user-friendly, e leggibile**. Questa รฉ la top motivazione per re-implementare modelli in ๐ค Transformers - cercare di creare nuove complesse tecnologie NLP accessibili a **chiunque**.
Riuscire a far girare il modello pretrained originale dalla repository ufficiale รฉ spesso il passo **piu arduo**. Dalla nostra esperienza, รฉ molto importante spendere un p' di tempo per diventare familiari con il codice base originale. Come test, prova a capire i seguenti punti:
- Dove si trovano i pretrained weights?
- Come caricare i pretrained weights nel modello corrispondente?
- Come girare un tokenizer independentemente dal modello?
- Prova a tracciare un singolo forward pass, cosicchรฉ potrai sapere che classi e funzioni sono richieste per un semplice forward pass. Di solito, dovrai reimplementare queste funzioni e basta
- Prova a localizzare i componenti importanti del modello: Dove si trova la classe del modello? Ci sono sotto classi nel modello *per esempio* EngoderModel, DecoderMOdel? Dove si trova il self-attention layer? Ci sono molteplici differenti layer di attention, *per esempio * *self-attention*, *cross-attention*...?
- Come puoi fare debug sul modello nell'ambiente originale della repo? Devi aggiungere dei *print* o puoi usare *ipdb* come debugger interattivo, o vabene anche un IDE efficiente per debug come PyCharm?
ร molto importante che prima di cominciare a trasferire il modello nuovo tu spenda tempo a fare debug del codice originale in maniera **efficiente**! Inoltre, ricorda che tutta la library รฉ open-soruce, quindi non temere di aprire issue o fare una pull request nella repo originale. Tutti coloro che mantengono la repository saranno piรบ che felici di avere qualcuno che guarda e gioca con i loro codici!
A questo punto, sta a te decidere quale ambiente per debug vuoi usare. Noi consilgiamo di evitare setup con GPU, che potrebbero costare assai, lavorare su una CPU puรณ essere un ottimo punto di partenza per indagare la repository originale e per cominciare a scrivere il codice per ๐ค Transformers. Solo alla fine, quando il modello รฉ stato portato con successo in ๐ค Transformers, allora si potrรก verificare il suo funzionamento su GPU.
In generale ci sono due possibili ambienti di debug per il testare il modello originale:
- [Jupyter notebooks](https://jupyter.org/) / [google colab](https://colab.research.google.com/notebooks/intro.ipynb)
- Scripts locali in Python
Il vantaggio dei Jupyter notebooks รฉ la possibilitร di eseguire cella per cella, il che puรฒ essere utile per decomporre tutte le componenti logiche, cosi da a vere un ciclo di debug piรน rapido, siccome si possono salvare i risultati da steps intermedi. Inoltre, i notebooks spesso sono molto facili da condividere con altri contributors, il che puรฒ essere molto utile se vuoi chiedere aiuto al team di Hugging Face. Se sei famigliare con Jupyter notebooks allora racommandiamo di lavorare in questa maniera.
Ovviamente se non siete abituati a lavorare con i notebook, questo puรฒ essere uno svantaggio nell'usare questa tecnologia, sprecando un sacco di tempo per setup e portare tutto al nuovo ambiente, siccome non potreste neanche usare dei tools di debug come `ipdb`.
Per ogni pratica code-base, รฉ sempre meglio come primo step caricare un **piccolo** checkpoint pretrained e cercare di riprodurre un singolo forward pass usando un vettore fittizio di IDs fatti da numeri interi. Un esempio per uno script simile, in pseudocodice รฉ:
```python
model = BrandNewBertModel.load_pretrained_checkpoint("/path/to/checkpoint/")
input_ids = [0, 4, 5, 2, 3, 7, 9] # vector of input ids
original_output = model.predict(input_ids)
```
Per quanto riguarda la strategia di debugging, si puรฒ scegliere tra:
- Decomporre il modello originario in piccole componenenti e testare ognuna di esse
- Decomporre il modello originario nel *tokenizer* originale e nel *modello* originale, testare un forward pass su questi,
e usare dei print statement o breakpoints intermedi per verificare
Ancora una volta, siete liberi di scegliere quale strategia sia ottimale per voi. Spesso una strategia รฉ piu
avvantaggiosa di un'altra, ma tutto dipende dall'code-base originario.
Se il code-base vi permette di decomporre il modello in piccole sub-componenenti, *per esempio* se il code-base
originario puรฒ essere facilmente testato in eager mode, allora vale la pena effettuare un debugging di questo genere.
Ricordate che ci sono dei vantaggi nel decidere di prendere la strada piu impegnativa sin da subito:
- negli stage piu finali, quando bisognerร comparare il modello originario all'implementazione in Hugging Face, potrete verificare
automaticamente ogni componente, individualmente, di modo che ci sia una corrispondenza 1:1
- avrete l'opportunitร di decomporre un problema molto grande in piccoli passi, cosรฌ da strutturare meglio il vostro lavoro
- separare il modello in componenti logiche vi aiuterร ad avere un'ottima overview sul design del modello, quindi una migliore
comprensione del modello stesso
- verso gli stage finali i test fatti componente per componente vi aiuterร ad essere sicuri di non andare avanti e indietro
nell'implementazione, cosรฌ da continuare la modifica del codice senza interruzione
Un ottimo esempio di come questo puรฒ essere fatto รฉ dato da [Lysandre](https://gist.github.com/LysandreJik/db4c948f6b4483960de5cbac598ad4ed)
per il modello ELECTRA
Tuttavia, se il code-base originale รฉ molto complesso o le componenti intermedie possono essere testate solo in tramite
compilazione, potrebbe richiedere parecchio tempo o addirittura essere impossibile separare il modello in piccole sotto-componenti.
Un buon esempio รฉ [MeshTensorFlow di T5](https://github.com/tensorflow/mesh/tree/master/mesh_tensorflow). Questa libreria
รฉ molto complessa e non offre un metodo semplice di decomposizione in sotto-componenti. Per simili librerie, potrete fare
affidamento ai print statements.
In ogni caso, indipendentemente da quale strategia scegliete, la procedura raccomandata รฉ di cominciare a fare debug dal
primo layer al layer finale.
ร consigliato recuperare gli output dai layers, tramite print o sotto-componenti, nel seguente ordine:
1. Recuperare gli IDs di input dati al modello
2. Recuperare i word embeddings
3. Recuperare l'input del primo Transformer layer
4. Recuperare l'output del primo Transformer layer
5. Recuperare l'output dei seguenti `n - 1` Transformer layers
6. Recuperare l'output dell'intero BrandNewBert Model
Gli IDs in input dovrebbero essere un arrary di interi, *per esempio* `input_ids = [0, 4, 4, 3, 2, 4, 1, 7, 19]`
Gli output dei seguenti layer di solito dovrebbero essere degli array di float multi-dimensionali come questo:
```
[[
[-0.1465, -0.6501, 0.1993, ..., 0.1451, 0.3430, 0.6024],
[-0.4417, -0.5920, 0.3450, ..., -0.3062, 0.6182, 0.7132],
[-0.5009, -0.7122, 0.4548, ..., -0.3662, 0.6091, 0.7648],
...,
[-0.5613, -0.6332, 0.4324, ..., -0.3792, 0.7372, 0.9288],
[-0.5416, -0.6345, 0.4180, ..., -0.3564, 0.6992, 0.9191],
[-0.5334, -0.6403, 0.4271, ..., -0.3339, 0.6533, 0.8694]]],
```
Ci aspettiamo che ogni modello aggiunto a ๐ค Transformers passi con successo un paio di test d'integrazione. Questo
significa che il modello originale e la sua implementazione in ๐ค Transformers abbiano lo stesso output con una precisione
di 0.001! Siccome รฉ normale che lo stesso esatto modello, scritto in librerie diverse, possa dare output leggermente
diversi, la tolleranza accettata รฉ 1e-3 (0.001). Ricordate che i due modelli devono dare output quasi identici. Dunque,
รฉ molto conveniente comparare gli output intermedi di ๐ค Transformers molteplici volte con gli output intermedi del
modello originale di *brand_new_bert*. Di seguito vi diamo alcuni consigli per avere un ambiente di debug il piu efficiente
possibile:
- Trovate la migliore strategia per fare debug dei risultati intermedi. Per esempio, รฉ la repository originale scritta in PyTorch?
Se si, molto probabilmente dovrete dedicare un po' di tempo per scrivere degli script piu lunghi, cosรฌ da decomporre il
modello originale in piccole sotto-componenti, in modo da poter recuperare i valori intermedi. Oppure, la repo originale
รฉ scritta in Tensorflow 1? Se รฉ cosรฌ dovrete fare affidamento ai print di Tensorflow [tf.print](https://www.tensorflow.org/api_docs/python/tf/print)
per avere i valori intermedi. Altro caso, la repo รฉ scritta in Jax? Allora assicuratevi che il modello non sia in **jit**
quanto testate il foward pass, *per esempio* controllate [questo link](https://github.com/google/jax/issues/196).
- Usate i piรน piccoli pretrained checkpoint che potete trovare. Piu piccolo รฉ il checkpoint, piu velocemente sarร il vostro
ciclo di debug. Non รฉ efficiente avere un pretrained model cosรฌ gigante che per il forward pass impieghi piu di 10 secondi.
Nel caso in cui i checkpoints siano molto grandi, e non si possa trovare di meglio, allora รฉ buona consuetudine ricorrere
a fare un dummy model nel nuovo ambiente, con weights inizializzati random e salvare quei weights per comprare la versione ๐ค Transformers
con il vostro modello
- Accertatevi di usare la via piu semplice per chiamare il forward pass nella repo originale. Sarebbe opportuno trovare
la funzione originaria che chiami **solo** un singolo forward pass, *per esempio* questa funzione spesso viene chiamata
`predict`, `evaluate`, `forward` o `__call__`. Siate sicuri di non fare debug su una funzione che chiami `forward` molteplici
volte, *per esempio* per generare testo, come `autoregressive_sample`, `generate`.
- Cercate di separare la tokenization dal forward pass del modello. Se la repo originaria mostra esempio dove potete dare
come input una stringa, provate a cercare dove nella forward call la stringa viene cambiata in input ids e cominciate il
debug da questo punto. Questo vi garantisce un ottimo punto di partenza per scrivere un piccolo script personale dove dare
gli input al modello, anziche delle stringhe in input.
- Assicuratevi che il debugging **non** sia in training mode. Spesso questo potra il modello a dare degli output random, per
via dei molteplici dropout layers. Assicuratevi che il forward pass nell'ambiente di debug sia **deterministico**, cosicche
i dropout non siano usati. Alternativamente, potete usare *transformers.utils.set_seed* se la vecchia e nuova implementazione
sono nello stesso framework.
La seguente sezione vi da ulteriori dettagli e accorgimenti su come potete fare tutto questo per *brand_new_bert*.
### 5.-14. Trasferire BrandNewBert in ๐ค Transformers
Allora cominciamo ad aggiungere un nuovo codice in ๐ค Transformers. Andate nel vostro fork clone di ๐ค Transformers:
```bash
cd transformers
```
Nel caso speciale in cui stiate aggiungendo un modello, la cui architettura sia identica a una di un modello giร esistente,
dovrete solo aggiugnere uno script di conversione, come descritto [qui](#write-a-conversion-script).
In questo caso, potete riutilizzare l'intera architettura del modello gia esistente.
Se questo non รฉ il caso, cominciamo con il generare un nuovo modello. Avrete due opzioni:
- `transformers-cli add-new-model-like` per aggiungere un nuovo modello come uno che gia esiste
- `transformers-cli add-new-model` per aggiungere un nuovo modello da un nostro template (questo assomigliera a BERT o Bart, in base al modello che selezionerete)
In entrambi i casi, l'output vi darร un questionario da riempire con informazioni basi sul modello. Il secondo comando richiede di installare
un `cookiecutter` - maggiori informazioni [qui](https://github.com/huggingface/transformers/tree/main/templates/adding_a_new_model).
**Aprire una Pull Request in main huggingface/transformers repo**
Prime di cominciare ad adattare il codice automaticamente generato, aprite una nuova PR come "Work in progress (WIP)",
*per esempio* "[WIP] Aggiungere *brand_new_bert*", cosicchรฉ il team di Hugging Face possa lavorare al vostro fianco nell'
integrare il modello in ๐ค Transformers.
Questi sarebbero gli step generali da seguire:
1. Creare un branch dal main branch con un nome descrittivo
```bash
git checkout -b add_brand_new_bert
```
2. Commit del codice automaticamente generato
```bash
git add .
git commit
```
3. Fare fetch e rebase del main esistente
```bash
git fetch upstream
git rebase upstream/main
```
4. Push dei cambiamenti al proprio account:
```bash
git push -u origin a-descriptive-name-for-my-changes
```
5. Una volte che siete soddisfatti dei nuovi cambiamenti, andate sulla webpage del vostro fork su GitHub. Cliccate "Pull request".
Assiuratevi di aggiungere alcuni membri di Hugging Face come reviewers, nel riguardo alla destra della pagina della PR, cosicche il team
Hugging Face verrร notificato anche per i futuri cambiamenti.
6. Cambiare la PR a draft, cliccando su "Convert to draft" alla destra della pagina della PR
Da quel punto in poi, ricordate di fare commit di ogni progresso e cambiamento, cosicche venga mostrato nella PR. Inoltre,
ricordatevi di tenere aggiornato il vostro lavoro con il main esistente:
```bash
git fetch upstream
git merge upstream/main
```
In generale, tutte le domande che avrete riguardo al modello o l'implementazione dovranno essere fatte nella vostra PR
e discusse/risolte nella PR stessa. In questa maniera, il team di Hugging Face sarร sempre notificato quando farete commit
di un nuovo codice o se avrete qualche domanda. ร molto utile indicare al team di Hugging Face il codice a cui fate riferimento
nella domanda, cosicche il team potra facilmente capire il problema o la domanda.
Per fare questo andate sulla tab "Files changed", dove potrete vedere tutti i vostri cambiamenti al codice, andate sulla linea
dove volete chiedere una domanda, e cliccate sul simbolo "+" per aggiungere un commento. Ogni volta che una domanda o problema
รฉ stato risolto, cliccate sul bottone "Resolve".
In questa stessa maniera, Hugging Face aprirร domande o commenti nel rivedere il vostro codice. Mi raccomando, chiedete piรน
domande possibili nella pagina della vostra PR. Se avete domande molto generali, non molto utili per il pubblico, siete liberi
di chiedere al team Hugging Face direttamente su slack o email.
**5. Adattare i codici per brand_new_bert**
Per prima cosa, ci focalizzeremo sul modello e non sui tokenizer. Tutto il codice relative dovrebbe trovarsi in
`src/transformers/models/brand_new_bert/modeling_brand_new_bert.py` e
`src/transformers/models/brand_new_bert/configuration_brand_new_bert.py`.
Ora potete finalmente cominciare il codice :). Il codice generato in
`src/transformers/models/brand_new_bert/modeling_brand_new_bert.py` avrร sia la stessa architettura di BERT se รฉ un
modello encoder-only o BART se รฉ encoder-decoder. A questo punto, ricordatevi cio che avete imparato all'inizio, riguardo
agli aspetti teorici del modello: *In che maniera il modello che sto implmementando รฉ diverso da BERT o BART?*. Implementare
questi cambi spesso vuol dire cambiare il layer *self-attention*, l'ordine dei layer di normalizzazione e cosรฌ via...
Ancora una volta ripetiamo, รฉ molto utile vedere architetture simili di modelli gia esistenti in Transformers per avere
un'idea migliore su come implementare il modello.
**Notate** che a questo punto non dovete avere subito un codice tutto corretto o pulito. Piuttosto, รฉ consigliato cominciare con un
codice poco pulito, con copia-incolla del codice originale in `src/transformers/models/brand_new_bert/modeling_brand_new_bert.py`
fino a che non avrete tutto il codice necessario. In base alla nostra esperienza, รฉ molto meglio aggiungere una prima bozza
del codice richiesto e poi correggere e migliorare iterativamente. L'unica cosa essenziale che deve funzionare qui รฉ la seguente
instanza:
```python
from transformers import BrandNewBertModel, BrandNewBertConfig
model = BrandNewBertModel(BrandNewBertConfig())
```
Questo comando creerร un modello con i parametri di default definiti in `BrandNewBergConfig()` e weights random. Questo garantisce
che `init()` di tutte le componenti funzioni correttamente.
**6. Scrivere uno script di conversione**
Il prossimo step รฉ scrivere uno script per convertire il checkpoint che avete usato per fare debug su *brand_new_berts* nella
repo originale in un checkpoint per la nuova implementazione di *brand_new_bert* in ๐ค Transformers. Non รฉ consigliato scrivere
lo script di conversione da zero, ma piuttosto cercate e guardate script gia esistenti in ๐ค Transformers, cosรฌ da trovarne
uno simile al vostro modello. Di solito basta fare una copia di uno script gia esistente e adattarlo al vostro caso.
Non esistate a chiedre al team di Hugging Face a riguardo.
- Se state convertendo un modello da TensorFlow a PyTorch, un ottimo inizio รฉ vedere [questo script di conversione per BERT](https://github.com/huggingface/transformers/blob/7acfa95afb8194f8f9c1f4d2c6028224dbed35a2/src/transformers/models/bert/modeling_bert.py#L91)
- Se state convertendo un modello da PyTorch a PyTorch, [lo script di conversione di BART puรฒ esservi utile](https://github.com/huggingface/transformers/blob/main/src/transformers/models/bart/convert_bart_original_pytorch_checkpoint_to_pytorch.py)
Qui di seguito spiegheremo come i modelli PyTorch salvano i weights per ogni layer e come i nomi dei layer sono definiti. In PyTorch,
il nomde del layer รฉ definito dal nome della class attribute che date al layer. Definiamo un modello dummy in PyTorch,
chiamato `SimpleModel`:
```python
from torch import nn
class SimpleModel(nn.Module):
def __init__(self):
super().__init__()
self.dense = nn.Linear(10, 10)
self.intermediate = nn.Linear(10, 10)
self.layer_norm = nn.LayerNorm(10)
```
Ora possiamo creare un'instanza di questa definizione di modo da inizializzare a random weights: `dense`, `intermediate`, `layer_norm`.
Possiamo usare print per vedere l'architettura del modello:
```python
model = SimpleModel()
print(model)
```
Da cui si ottiene:
```
SimpleModel(
(dense): Linear(in_features=10, out_features=10, bias=True)
(intermediate): Linear(in_features=10, out_features=10, bias=True)
(layer_norm): LayerNorm((10,), eps=1e-05, elementwise_affine=True)
)
```
Si puรฒ vedere come i nomi dei layers siano definiti dal nome della class attribute in PyTorch. I valori dei weights di uno
specifico layer possono essere visualizzati:
```python
print(model.dense.weight.data)
```
ad esempio:
```
tensor([[-0.0818, 0.2207, -0.0749, -0.0030, 0.0045, -0.1569, -0.1598, 0.0212,
-0.2077, 0.2157],
[ 0.1044, 0.0201, 0.0990, 0.2482, 0.3116, 0.2509, 0.2866, -0.2190,
0.2166, -0.0212],
[-0.2000, 0.1107, -0.1999, -0.3119, 0.1559, 0.0993, 0.1776, -0.1950,
-0.1023, -0.0447],
[-0.0888, -0.1092, 0.2281, 0.0336, 0.1817, -0.0115, 0.2096, 0.1415,
-0.1876, -0.2467],
[ 0.2208, -0.2352, -0.1426, -0.2636, -0.2889, -0.2061, -0.2849, -0.0465,
0.2577, 0.0402],
[ 0.1502, 0.2465, 0.2566, 0.0693, 0.2352, -0.0530, 0.1859, -0.0604,
0.2132, 0.1680],
[ 0.1733, -0.2407, -0.1721, 0.1484, 0.0358, -0.0633, -0.0721, -0.0090,
0.2707, -0.2509],
[-0.1173, 0.1561, 0.2945, 0.0595, -0.1996, 0.2988, -0.0802, 0.0407,
0.1829, -0.1568],
[-0.1164, -0.2228, -0.0403, 0.0428, 0.1339, 0.0047, 0.1967, 0.2923,
0.0333, -0.0536],
[-0.1492, -0.1616, 0.1057, 0.1950, -0.2807, -0.2710, -0.1586, 0.0739,
0.2220, 0.2358]]).
```
Nello script di conversione, dovreste riempire quei valori di inizializzazione random con gli stessi weights del corrispondente
layer nel checkpoint. *Per esempio*
```python
# retrieve matching layer weights, e.g. by
# recursive algorithm
layer_name = "dense"
pretrained_weight = array_of_dense_layer
model_pointer = getattr(model, "dense")
model_pointer.weight.data = torch.from_numpy(pretrained_weight)
```
Cosรฌ facendo, dovete verificare che ogni inizializzazione random di un peso del modello PyTorch e il suo corrispondente peso nel pretrained checkpoint
siano esattamente gli stessi e uguali in **dimensione/shape e nome**. Per fare questo, รฉ **necessario** aggiungere un `assert`
per la dimensione/shape e nome:
```python
assert (
model_pointer.weight.shape == pretrained_weight.shape
), f"Pointer shape of random weight {model_pointer.shape} and array shape of checkpoint weight {pretrained_weight.shape} mismatched"
```
Inoltre, dovrete fare il print sia dei nomi che dei weights per essere sicuri che siano gli stessi:
```python
logger.info(f"Initialize PyTorch weight {layer_name} from {pretrained_weight.name}")
```
Se la dimensione o il nome non sono uguali, probabilmente avete sbagliato ad assegnare il peso nel checkpoint o nel layer costrutture di
๐ค Transformers.
Una dimensione sbagliata puรฒ essere dovuta ad un errore nei parameteri in `BrandNewBertConfig()`. Tuttavia, puรฒ essere anche
che l'implementazione del layer in PyTorch richieda di fare una transposizione della matrice dei weights.
Infine, controllate **tutti** che tutti i weights inizializzati e fate print di tutti i weights del checkpoint che non sono stati
usati per l'inizializzazione, di modo da essere sicuri che il modello sia correttamente convertito. ร normale che ci siano
errori nel test di conversione, fai per un errore in `BrandNewBertConfig()`, o un errore nell'architettura in ๐ค Transformers,
o un bug in `init()`.
Questo step dev'essere fatto tramite iterazioni fino a che non si raggiungano gli stessi valori per i weights. Una volta che
il checkpoint รฉ stato correttamente caricato in ๐ค Transformers, potete salvare il modello in una cartella di vostra scelta
`/path/to/converted/checkpoint/folder` che contenga sia
`pytorch_model.bin` che `config.json`:
```python
model.save_pretrained("/path/to/converted/checkpoint/folder")
```
**7. Implementare il forward pass**
Una volta che i weights pretrained sono stati correttamente caricati in ๐ค Transformers, dovrete assicurarvi che il forward pass
sia correttamente implementato. [Qui](#provare-un-pretrained-checkpoint-usando-la-repo-originale), avete give creato e provato
uno script che testi il forward pass del modello usando la repo originaria. Ora dovrete fare lo stesso con uno script analogo
usando l'implementazione in ๐ค Transformers anzichรฉ l'originale. Piu o meno lo script dovrebbe essere:
```python
model = BrandNewBertModel.from_pretrained("/path/to/converted/checkpoint/folder")
input_ids = [0, 4, 4, 3, 2, 4, 1, 7, 19]
output = model(input_ids).last_hidden_states
```
Di solito l'output da ๐ค Transformers non รฉ uguale uguale all'output originario, sopratto la prima volta. Non vi abbattete -
รฉ normale! Prima di tutto assicuratevi che non ci siano errori o che non vengano segnalati degli errori nella forward pass.
Spesso capita che ci siano dimensioni sbagliate o data type sbagliati, *ad esempio* `torch.long` anziche `torch.float32`.
Non esistate a chiedere al team Hugging Face!
Nella parte finale assicuratevi che l'implementazione ๐ค Transformers funzioni correttamente cosi da testare che gli output
siano equivalenti a una precisione di `1e-3`. Controllate che `outputs.shape` siano le stesse tra ๐ค Transformers e l'implementazione
originaria. Poi, controllate che i valori in output siano identici. Questa รฉ sicuramente la parte piรน difficile, qui una serie
di errori comuni quando gli output non sono uguali:
- Alcuni layers non sono stati aggiunti, *ad esempio* un *activation* layer non รฉ stato aggiunto, o ci si รฉ scordati di una connessione
- La matrice del word embedding non รฉ stata ripareggiata
- Ci sono degli embeddings posizionali sbagliati perchรฉ l'implementazione originaria ha un offset
- Il dropout รฉ in azione durante il forward pass. Per sistemare questo errore controllate che *model.training = False* e che
il dropout non sia stato attivato nel forward pass, * per esempio * passate *self.training* a [PyTorch's functional dropout](https://pytorch.org/docs/stable/nn.functional.html?highlight=dropout#torch.nn.functional.dropout)
La miglior maniera per sistemare il problema รฉ di vedere all'implementazione originaria del forward pass e in ๐ค Transformers
fianco a fianco e vedere se ci sono delle differenze. In teoria, con debug e print degli output intermedie di entrambe le
implementazioni nel forward pass nell'esatta posizione del network dovrebbe aiutarvi a vedere dove ci sono differenze tra
i due frameworks. Come prima mossa controllate che `input_ids` siano identici in entrambi gli scripts. Da lรฌ andate fino
all'ultimo layer. Potrete notare una differenza tra le due implementazioni a quel punto.
Una volta che lo stesso output รฉ stato ragguingi, verificate gli output con `torch.allclose(original_output, output, atol=1e-3)`.
A questo punto se รฉ tutto a posto: complimenti! Le parti seguenti saranno una passeggiata ๐.
**8. Aggiungere i test necessari per il modello**
A questo punto avete aggiunto con successo il vostro nuovo modello. Tuttavia, รฉ molto probabile che il modello non sia
del tutto ok con il design richiesto. Per essere sicuri che l'implementazione sia consona e compatibile con ๐ค Transformers รฉ
necessario implementare dei tests. Il Cookiecutter dovrebbe fornire automaticamente dei file per test per il vostro modello,
di solito nella folder `tests/test_modeling_brand_new_bert.py`. Provate questo per verificare l'ok nei test piu comuni:
```bash
pytest tests/test_modeling_brand_new_bert.py
```
Una volta sistemati i test comuni, bisogna assicurarsi che il vostro lavoro sia correttamente testato cosicchรจ:
- a) La community puo capire in maniera semplice il vostro lavoro controllando tests specifici del modello *brand_new_bert*,
- b) Implementazioni future del vostro modello non rompano alcune feature importante del modello.
Per prima cosa agguingete dei test d'integrazione. Questi sono essenziali perche fanno la stessa funzione degli scripts di
debug usati precedentemente. Un template per questi tests esiste gia nel Cookiecutter ed รฉ sotto il nome di `BrandNewBertModelIntegrationTests`,
voi dovrete solo completarlo. Una volta che questi tests sono OK, provate:
```bash
RUN_SLOW=1 pytest -sv tests/test_modeling_brand_new_bert.py::BrandNewBertModelIntegrationTests
```
<Tip>
Nel caso siate su Windows, sostituite `RUN_SLOW=1` con `SET RUN_SLOW=1`
</Tip>
Di seguito, tutte le features che sono utili e necessarire per *brand_new_bert* devono essere testate in test separati,
contenuti in `BrandNewBertModelTester`/ `BrandNewBertModelTest`. spesso la gente si scorda questi test, ma ricordate che sono utili per:
- Aiuta gli utenti a capire il vostro codice meglio, richiamando l'attenzione su queste nuove features
- Developers e contributors futuri potranno velocemente testare nuove implementazioni del modello testanto questi casi speciali.
**9. Implementare il tokenizer**
A questo punto avremo bisogno un tokenizer per *brand_new_bert*. Di solito il tokenizer รฉ uguale ad altri modelli in ๐ค Transformers.
ร importante che troviate il file con il tokenizer originale e che lo carichiate in ๐ค Transformers.
Per controllare che il tokenizer funzioni in modo corretto, create uno script nella repo originaria che riceva come input
una stringa e ritorni gli `input_ids`. Piu o meno questo potrebbe essere il codice:
```python
input_str = "This is a long example input string containing special characters .$?-, numbers 2872 234 12 and words."
model = BrandNewBertModel.load_pretrained_checkpoint("/path/to/checkpoint/")
input_ids = model.tokenize(input_str)
```
Potrebbe richiedere un po' di tempo, ma guardate ancora alla repo originaria per trovare la funzione corretta del tokenizer.
A volte capita di dover riscrivere il tokenizer nella repo originaria, di modo da avere come output gli `input_ids`.
A quel punto uno script analogo รฉ necessario in ๐ค Transformers:
```python
from transformers import BrandNewBertTokenizer
input_str = "This is a long example input string containing special characters .$?-, numbers 2872 234 12 and words."
tokenizer = BrandNewBertTokenizer.from_pretrained("/path/to/tokenizer/folder/")
input_ids = tokenizer(input_str).input_ids
```
Una volta che `input_ids` sono uguali, bisogna aggiungere un test per il tokenizer.
Il file test per tokenizer di *brand_new_brand* dovrebbe avere un paio di hard-coded test d'integrazione.
**10. Test end-to-end**
Ora che avete il tokenizer, dovrete aggiungere dei test d'integrazione per l'intero workflow in `tests/test_modeling_brand_new_bert.py` in ๐ค Transformer.
Questi test devono mostrare che un significante campione text-to-text funzioni come ci si aspetta nell'implementazione di ๐ค Transformers.
*Per esempio* potreste usare dei source-to-target-translation, o un sommario di un articolo, o un domanda-risposta e cosi via.
Se nessuno dei checkpoints รฉ stato ultra parametrizzato per task simili, allora i tests per il modello sono piu che sufficienti.
Nello step finale dovete assicurarvi che il modello sia totalmente funzionale, e consigliamo anche di provare a testare su GPU.
Puo succedere che ci si scordi un `.to(self.device)` ad esempio. Se non avete accesso a GPU, il team Hugging Face puo provvedere
a testare questo aspetto per voi.
**11. Aggiungere una Docstring**
Siete quasi alla fine! L'ultima cosa rimasta รฉ avere una bella docstring e una pagina doc. Il Cookiecutter dovrebbe provvedere giร
un template chiamato `docs/source/model_doc/brand_new_bert.rst`, che dovrete compilare. La prima cosa che un utente farร
per usare il vostro modello sarร dare una bella lettura al doc. Quindi proponete una documentazione chiara e concisa. ร molto
utile per la community avere anche delle *Tips* per mostrare come il modello puo' essere usato. Non esitate a chiedere a Hugging Face
riguardo alle docstirng.
Quindi, assicuratevi che la docstring sia stata aggiunta a `src/transformers/models/brand_new_bert/modeling_brand_new_bert.py`.
Assicuratevi che la docstring sia corretta e che includa tutti i necessari input e output. Abbiamo una guida dettagliata per
scrivere la documentazione e docstring.
**Rifattorizzare il codice**
Perfetto! Ora che abbiamo tutto per *brand_new_bert* controllate che lo stile del codice sia ok:
```bash
make style
```
E che il codice passi i quality check:
```bash
make quality
```
A volte capita che manchino delle informazioninella docstring o alcuni nomi sbagliati, questo farร fallire i tests sopra.
Ripetiamo: chiedete pure a Hugging Face, saremo lieti di aiutarvi.
Per ultimo, fare del refactoring del codice una volta che รฉ stato creato.
Avete finito con il codice, congratulazioni! ๐ Siete fantasticiiiiiii! ๐
**12. Caricare il modello sul model hub**
In questa ultima parte dovrete convertire e caricare il modello, con tutti i checkpoints, nel model hub e aggiungere una
model card per ogni checkpoint caricato. Leggete la nostra guida [Model sharing and uploading Page](model_sharing) per
avere familiaritร con l'hub. Di solito in questa parte lavorate a fianco di Hugging face per decidere un nome che sia ok
per ogni checkpoint, per ottenere i permessi necessari per caricare il modello nell'organizzazione dell'autore di *brand_new_bert*.
Il metodo `push_to_hub`, presente in tutti i modelli `transformers`, รฉ una maniera rapida e indolore per caricare il vostro checkpoint sull'hub:
```python
brand_new_bert.push_to_hub(
repo_path_or_name="brand_new_bert",
# Uncomment the following line to push to an organization
# organization="<ORGANIZATION>",
commit_message="Add model",
use_temp_dir=True,
)
```
Vale la pena spendere un po' di tempo per creare una model card ad-hoc per ogni checkpoint. Le model cards dovrebbero
suggerire le caratteristiche specifiche del checkpoint, *per esempio* su che dataset il checkpoint รฉ stato pretrained o fine-tuned.
O che su che genere di task il modello lavoro? E anche buona pratica includere del codice su come usare il modello correttamente.
**13. (Opzionale) Aggiungere un notebook**
ร molto utile aggiungere un notebook, che dimostri in dettaglio come *brand_new_bert* si utilizzi per fare inferenza e/o
fine-tuned su specifiche task. Non รฉ una cosa obbligatoria da avere nella vostra PR, ma รฉ molto utile per la community.
**14. Sottomettere la PR**
L'ultimissimo step! Ovvero il merge della PR nel main. Di solito il team Hugging face a questo punto vi avrร gia aiutato,
ma รฉ ok prendere un po' di tempo per pulire la descirzione e commenti nel codice.
### Condividete il vostro lavoro!!
ร ora tempo di prendere un po' di credito dalla communitร per il vostro lavoro! Caricare e implementare un nuovo modello
รฉ un grandissimo contributo per Transformers e l'intera community NLP. Il codice e la conversione dei modelli pre-trained sara
sicuramente utilizzato da centinaia o migliaia di sviluppatori e ricercatori. Siate fieri e orgogliosi di condividere il vostro
traguardo con l'intera community :)
** Avete create un altro modello che รฉ super facile da usare per tutti quanti nella community! ๐คฏ**
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/run_scripts.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Addestramento con script
Insieme ai [notebooks](./noteboks/README) ๐ค Transformers, ci sono anche esempi di script che dimostrano come addestrare un modello per un task con [PyTorch](https://github.com/huggingface/transformers/tree/main/examples/pytorch), [TensorFlow](https://github.com/huggingface/transformers/tree/main/examples/tensorflow), o [JAX/Flax](https://github.com/huggingface/transformers/tree/main/examples/flax).
Troverai anche script che abbiamo usato nei nostri [progetti di ricerca](https://github.com/huggingface/transformers/tree/main/examples/research_projects) e [precedenti esempi](https://github.com/huggingface/transformers/tree/main/examples/legacy) a cui contribuisce per lo piรน la comunitร . Questi script non sono attivamente mantenuti e richiedono una specifica versione di ๐ค Transformers che sarร molto probabilmente incompatibile con l'ultima versione della libreria.
Non รจ dato per scontato che gli script di esempio funzionino senza apportare modifiche per ogni problema, bensรฌ potrebbe essere necessario adattare lo script al tuo caso specifico. Per aiutarti in ciรฒ, la maggioranza degli script espone le modalitร di pre-processamento dei dati, consentendoti di modificare lo script come preferisci.
Per qualsiasi feature che vorresti implementare in uno script d'esempio, per favore discutine nel [forum](https://discuss.huggingface.co/) o in un'[issue](https://github.com/huggingface/transformers/issues) prima di inviare una Pull Request. Mentre accogliamo con piacere la correzione di bug, รจ piรน improbabile che faremo la stessa con una PR che aggiunge funzionalitร sacrificando la leggibilitร .
Questa guida ti mostrerร come eseguire uno script di esempio relativo al task di summarization in [PyTorch](https://github.com/huggingface/transformers/tree/main/examples/pytorch/summarization) e [TensorFlow](https://github.com/huggingface/transformers/tree/main/examples/tensorflow/summarization). Tutti gli esempi funzioneranno con entrambi i framework a meno che non sia specificato altrimenti.
## Installazione
Per eseguire con successo l'ultima versione degli script di esempio, devi **installare ๐ค Transformers dalla fonte** in un nuovo ambiente virtuale:
```bash
git clone https://github.com/huggingface/transformers
cd transformers
pip install .
```
Per le precedenti versioni degli script di esempio, clicca sul pulsante di seguito:
<details>
<summary>Esempi per versioni precedenti di ๐ค Transformers</summary>
<ul>
<li><a href="https://github.com/huggingface/transformers/tree/v4.5.1/examples">v4.5.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v4.4.2/examples">v4.4.2</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v4.3.3/examples">v4.3.3</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v4.2.2/examples">v4.2.2</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v4.1.1/examples">v4.1.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v4.0.1/examples">v4.0.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.5.1/examples">v3.5.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.4.0/examples">v3.4.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.3.1/examples">v3.3.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.2.0/examples">v3.2.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.1.0/examples">v3.1.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v3.0.2/examples">v3.0.2</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.11.0/examples">v2.11.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.10.0/examples">v2.10.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.9.1/examples">v2.9.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.8.0/examples">v2.8.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.7.0/examples">v2.7.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.6.0/examples">v2.6.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.5.1/examples">v2.5.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.4.0/examples">v2.4.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.3.0/examples">v2.3.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.2.0/examples">v2.2.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.1.0/examples">v2.1.1</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v2.0.0/examples">v2.0.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v1.2.0/examples">v1.2.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v1.1.0/examples">v1.1.0</a></li>
<li><a href="https://github.com/huggingface/transformers/tree/v1.0.0/examples">v1.0.0</a></li>
</ul>
</details>
Successivamente, cambia la tua attuale copia di ๐ค Transformers specificandone la versione, ad esempio v3.5.1:
```bash
git checkout tags/v3.5.1
```
Dopo aver configurato correttamente la versione della libreria, naviga nella cartella degli esempi di tua scelta e installa i requisiti:
```bash
pip install -r requirements.txt
```
## Esegui uno script
<frameworkcontent>
<pt>
Lo script di esempio scarica e pre-processa un dataset dalla libreria ๐ค [Datasets](https://huggingface.co/docs/datasets/). Successivamente, lo script esegue il fine-tuning su un dataset usando il [Trainer](https://huggingface.co/docs/transformers/main_classes/trainer) su un'architettura che supporta la summarization. Il seguente esempio mostra come eseguire il fine-tuning di [T5-small](https://huggingface.co/t5-small) sul dataset [CNN/DailyMail](https://huggingface.co/datasets/cnn_dailymail). Il modello T5 richiede un parametro addizionale `source_prefix` a causa del modo in cui รจ stato addestrato. Questo prefisso permette a T5 di sapere che si tratta di un task di summarization.
```bash
python examples/pytorch/summarization/run_summarization.py \
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--predict_with_generate
```
</pt>
<tf>
Lo script di esempio scarica e pre-processa un dataset dalla libreria ๐ค [Datasets](https://huggingface.co/docs/datasets/). Successivamente, lo script esegue il fine-tuning su un dataset usando Keras su un'architettura che supporta la summarization. Il seguente esempio mostra come eseguire il fine-tuning di [T5-small](https://huggingface.co/t5-small) sul dataset [CNN/DailyMail](https://huggingface.co/datasets/cnn_dailymail). Il modello T5 richiede un parametro addizionale `source_prefix` a causa del modo in cui รจ stato addestrato. Questo prefisso permette a T5 di sapere che si tratta di un task di summarization.
```bash
python examples/tensorflow/summarization/run_summarization.py \
--model_name_or_path t5-small \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size 8 \
--per_device_eval_batch_size 16 \
--num_train_epochs 3 \
--do_train \
--do_eval
```
</tf>
</frameworkcontent>
## Addestramento distribuito e precisione mista
Il [Trainer](https://huggingface.co/docs/transformers/main_classes/trainer) supporta l'addestramento distribuito e la precisione mista, che significa che puoi anche usarla in uno script. Per abilitare entrambe le funzionalitร :
- Aggiunto l'argomento `fp16` per abilitare la precisione mista.
- Imposta un numero di GPU da usare con l'argomento `nproc_per_node`.
```bash
torchrun \
--nproc_per_node 8 pytorch/summarization/run_summarization.py \
--fp16 \
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--predict_with_generate
```
Gli script TensorFlow utilizzano una [`MirroredStrategy`](https://www.tensorflow.org/guide/distributed_training#mirroredstrategy) per il training distribuito e non devi aggiungere alcun argomento addizionale allo script di training. Lo script TensorFlow userร multiple GPU in modo predefinito se quest'ultime sono disponibili:
## Esegui uno script su TPU
<frameworkcontent>
<pt>
Le Tensor Processing Units (TPU) sono state progettate per migliorare le prestazioni. PyTorch supporta le TPU con il compilatore per deep learning [XLA](https://www.tensorflow.org/xla) (guarda [questo link](https://github.com/pytorch/xla/blob/master/README.md) per maggiori dettagli). Per usare una TPU, avvia lo script `xla_spawn.py` e usa l'argomento `num_cores` per impostare il numero di core TPU che intendi usare.
```bash
python xla_spawn.py --num_cores 8 \
summarization/run_summarization.py \
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--predict_with_generate
```
</pt>
<tf>
Le Tensor Processing Units (TPU) sono state progettate per migliorare le prestazioni. Gli script TensorFlow utilizzano una [`TPUStrategy`](https://www.tensorflow.org/guide/distributed_training#tpustrategy) per eseguire l'addestramento su TPU. Per usare una TPU, passa il nome della risorsa TPU all'argomento `tpu`.
```bash
python run_summarization.py \
--tpu name_of_tpu_resource \
--model_name_or_path t5-small \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size 8 \
--per_device_eval_batch_size 16 \
--num_train_epochs 3 \
--do_train \
--do_eval
```
</tf>
</frameworkcontent>
## Esegui uno script con ๐ค Accelerate
๐ค [Accelerate](https://huggingface.co/docs/accelerate) รจ una libreria compatibile solo con PyTorch che offre un metodo unificato per addestrare modelli su diverse tipologie di configurazioni (CPU, multiple GPU, TPU) mantenendo una completa visibilitร rispetto al ciclo di training di PyTorch. Assicurati di aver effettuato l'installazione di ๐ค Accelerate, nel caso non lo avessi fatto:
> Nota: dato che Accelerate รจ in rapido sviluppo, รจ necessario installare la versione proveniente da git per eseguire gli script:
```bash
pip install git+https://github.com/huggingface/accelerate
```
Invece che usare lo script `run_summarization.py`, devi usare lo script `run_summarization_no_trainer.py`. Gli script supportati in ๐ค Accelerate avranno un file chiamato `task_no_trainer.py` nella rispettiva cartella. Per iniziare, esegui il seguente comando per creare e salvare un file di configurazione:
```bash
accelerate config
```
Testa la tua configurazione per assicurarti della sua correttezza:
```bash
accelerate test
```
Ora sei pronto per avviare l'addestramento:
```bash
accelerate launch run_summarization_no_trainer.py \
--model_name_or_path t5-small \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir ~/tmp/tst-summarization
```
## Uso di un dataset personalizzato
Lo script di summarization supporta dataset personalizzati purchรฉ siano file CSV o JSON Line. Quando usi il tuo dataset, devi specificare diversi argomenti aggiuntivi:
- `train_file` e `validation_file` specificano dove si trovano i file di addestramento e validazione.
- `text_column` รจ il file di input da riassumere.
- `summary_column` รจ il file di destinazione per l'output.
Uno script di summarization usando un dataset personalizzato sarebbe simile a questo:
```bash
python examples/pytorch/summarization/run_summarization.py \
--model_name_or_path t5-small \
--do_train \
--do_eval \
--train_file path_to_csv_or_jsonlines_file \
--validation_file path_to_csv_or_jsonlines_file \
--text_column text_column_name \
--summary_column summary_column_name \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--overwrite_output_dir \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--predict_with_generate
```
## Testare uno script
ร spesso una buona idea avviare il tuo script su un numero inferiore di esempi tratti dal dataset, per assicurarti che tutto funzioni come previsto prima di eseguire lo script sull'intero dataset, che potrebbe necessitare di ore. Usa i seguenti argomenti per limitare il dataset ad un massimo numero di esempi:
- `max_train_samples`
- `max_eval_samples`
- `max_predict_samples`
```bash
python examples/pytorch/summarization/run_summarization.py \
--model_name_or_path t5-small \
--max_train_samples 50 \
--max_eval_samples 50 \
--max_predict_samples 50 \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--predict_with_generate
```
Non tutti gli esempi di script supportano l'argomento `max_predict_samples`. Se non sei sicuro circa il supporto di questo argomento da parte del tuo script, aggiungi l'argomento `-h` per controllare:
```bash
examples/pytorch/summarization/run_summarization.py -h
```
## Riavviare addestramento da un checkpoint
Un'altra utile opzione รจ riavviare un addestramento da un checkpoint precedente. Questo garantirร che tu possa riprendere da dove hai interrotto senza ricominciare se l'addestramento viene interrotto. Ci sono due metodi per riavviare l'addestramento da un checkpoint:
Il primo metodo usa l'argomento `output_dir previous_output_dir` per riavviare l'addestramento dall'ultima versione del checkpoint contenuto in `output_dir`. In questo caso, dovresti rimuovere `overwrite_output_dir`:
```bash
python examples/pytorch/summarization/run_summarization.py
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--output_dir previous_output_dir \
--predict_with_generate
```
Il secondo metodo usa l'argomento `resume_from_checkpoint path_to_specific_checkpoint` per riavviare un addestramento da una specifica cartella di checkpoint.
```bash
python examples/pytorch/summarization/run_summarization.py
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--resume_from_checkpoint path_to_specific_checkpoint \
--predict_with_generate
```
## Condividi il tuo modello
Tutti gli script possono caricare il tuo modello finale al [Model Hub](https://huggingface.co/models). Prima di iniziare, assicurati di aver effettuato l'accesso su Hugging Face:
```bash
huggingface-cli login
```
Poi, aggiungi l'argomento `push_to_hub` allo script. Questo argomento consentirร di creare un repository con il tuo username Hugging Face e la cartella specificata in `output_dir`.
Per dare uno specifico nome al repository, usa l'argomento `push_to_hub_model_id`. Il repository verrร automaticamente elencata sotto al tuo namespace.
Il seguente esempio mostra come caricare un modello specificando il nome del repository:
```bash
python examples/pytorch/summarization/run_summarization.py
--model_name_or_path t5-small \
--do_train \
--do_eval \
--dataset_name cnn_dailymail \
--dataset_config "3.0.0" \
--source_prefix "summarize: " \
--push_to_hub \
--push_to_hub_model_id finetuned-t5-cnn_dailymail \
--output_dir /tmp/tst-summarization \
--per_device_train_batch_size=4 \
--per_device_eval_batch_size=4 \
--overwrite_output_dir \
--predict_with_generate
```
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/perf_infer_special.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Inferenza su Hardware Specializzato
Questo documento sarร completato a breve con la documentazione per l'inferenza su hardware specializzato. Nel frattempo puoi controllare [la guida per fare inferenza sulle CPU](perf_infer_cpu).
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/_config.py
|
# docstyle-ignore
INSTALL_CONTENT = """
# Installazione di Transformers
! pip install transformers datasets
# Per installare dalla fonte invece dell'ultima versione rilasciata, commenta il comando sopra e
# rimuovi la modalitร commento al comando seguente.
# ! pip install git+https://github.com/huggingface/transformers.git
"""
notebook_first_cells = [{"type": "code", "content": INSTALL_CONTENT}]
black_avoid_patterns = {
"{processor_class}": "FakeProcessorClass",
"{model_class}": "FakeModelClass",
"{object_class}": "FakeObjectClass",
}
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/serialization.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Esporta modelli ๐ค Transformers
Se devi implementare ๐ค modelli Transformers in ambienti di produzione, noi
consigliamo di esportarli in un formato serializzato che puรฒ essere caricato ed eseguito
su runtime e hardware specializzati. In questa guida ti mostreremo come farlo
esporta ๐ค Modelli Transformers in due formati ampiamente utilizzati: ONNX e TorchScript.
Una volta esportato, un modello puรฒ essere ottimizato per l'inferenza tramite tecniche come
la quantizzazione e soppressione. Se sei interessato a ottimizzare i tuoi modelli per l'esecuzione
con la massima efficienza, dai un'occhiata a [๐ค Optimum
library](https://github.com/huggingface/optimum).
## ONNX
Il progetto [ONNX (Open Neural Network eXchange)](http://onnx.ai) Il progetto onnx รจ un open
standard che definisce un insieme comune di operatori e un formato di file comune a
rappresentano modelli di deep learning in un'ampia varietร di framework, tra cui
PyTorch e TensorFlow. Quando un modello viene esportato nel formato ONNX, questi
operatori sono usati per costruire un grafico computazionale (often called an
_intermediate representation_) che rappresenta il flusso di dati attraverso la
rete neurale.
Esponendo un grafico con operatori e tipi di dati standardizzati, ONNX rende
piรน facile passare da un framework all'altro. Ad esempio, un modello allenato in PyTorch puรฒ
essere esportato in formato ONNX e quindi importato in TensorFlow (e viceversa).
๐ค Transformers fornisce un pacchetto `transformers.onnx` che ti consente di
convertire i checkpoint del modello in un grafico ONNX sfruttando gli oggetti di configurazione.
Questi oggetti di configurazione sono giร pronti per una serie di architetture di modelli,
e sono progettati per essere facilmente estensibili ad altre architetture.
Le configurazioni pronte includono le seguenti architetture:
<!--This table is automatically generated by `make fix-copies`, do not fill manually!-->
- ALBERT
- BART
- BEiT
- BERT
- BigBird
- BigBird-Pegasus
- Blenderbot
- BlenderbotSmall
- CamemBERT
- ConvBERT
- Data2VecText
- Data2VecVision
- DeiT
- DistilBERT
- ELECTRA
- FlauBERT
- GPT Neo
- GPT-J
- I-BERT
- LayoutLM
- M2M100
- Marian
- mBART
- MobileBERT
- OpenAI GPT-2
- Perceiver
- PLBart
- RoBERTa
- RoFormer
- SqueezeBERT
- T5
- ViT
- XLM
- XLM-RoBERTa
- XLM-RoBERTa-XL
Nelle prossime due sezioni, ti mostreremo come:
* Esporta un modello supportato usando il pacchetto `transformers.onnx`.
* Esporta un modello personalizzato per un'architettura non supportata.
### Esportazione di un modello in ONNX
Per esportare un modello ๐ค Transformers in ONNX, dovrai prima installarne alcune
dipendenze extra:
```bash
pip install transformers[onnx]
```
Il pacchetto `transformers.onnx` puรฒ essere usato come modulo Python:
```bash
python -m transformers.onnx --help
usage: Hugging Face Transformers ONNX exporter [-h] -m MODEL [--feature {causal-lm, ...}] [--opset OPSET] [--atol ATOL] output
positional arguments:
output Path indicating where to store generated ONNX model.
optional arguments:
-h, --help show this help message and exit
-m MODEL, --model MODEL
Model ID on huggingface.co or path on disk to load model from.
--feature {causal-lm, ...}
The type of features to export the model with.
--opset OPSET ONNX opset version to export the model with.
--atol ATOL Absolute difference tolerance when validating the model.
```
L'esportazione di un checkpoint utilizzando una configurazione giร pronta puรฒ essere eseguita come segue:
```bash
python -m transformers.onnx --model=distilbert-base-uncased onnx/
```
che dovrebbe mostrare i seguenti log:
```bash
Validating ONNX model...
-[โ] ONNX model output names match reference model ({'last_hidden_state'})
- Validating ONNX Model output "last_hidden_state":
-[โ] (2, 8, 768) matches (2, 8, 768)
-[โ] all values close (atol: 1e-05)
All good, model saved at: onnx/model.onnx
```
Questo esporta un grafico ONNX del checkpoint definito dall'argomento `--model`.
In questo esempio รจ `distilbert-base-uncased`, ma puรฒ essere qualsiasi checkpoint
Hugging Face Hub o uno memorizzato localmente.
Il file risultante `model.onnx` puรฒ quindi essere eseguito su uno dei [tanti
acceleratori](https://onnx.ai/supported-tools.html#deployModel) che supportano il
lo standard ONNX. Ad esempio, possiamo caricare ed eseguire il modello con [ONNX
Runtime](https://onnxruntime.ai/) come segue:
```python
>>> from transformers import AutoTokenizer
>>> from onnxruntime import InferenceSession
>>> tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased")
>>> session = InferenceSession("onnx/model.onnx")
>>> # ONNX Runtime expects NumPy arrays as input
>>> inputs = tokenizer("Using DistilBERT with ONNX Runtime!", return_tensors="np")
>>> outputs = session.run(output_names=["last_hidden_state"], input_feed=dict(inputs))
```
I nomi di output richiesti (cioรจ `["last_hidden_state"]`) possono essere ottenuti
dando un'occhiata alla configurazione ONNX di ogni modello. Ad esempio, per
DistilBERT abbiamo:
```python
>>> from transformers.models.distilbert import DistilBertConfig, DistilBertOnnxConfig
>>> config = DistilBertConfig()
>>> onnx_config = DistilBertOnnxConfig(config)
>>> print(list(onnx_config.outputs.keys()))
["last_hidden_state"]
```
Il processo รจ identico per i checkpoint TensorFlow sull'hub. Ad esempio, noi
possiamo esportare un checkpoint TensorFlow puro da [Keras
organizzazione](https://huggingface.co/keras-io) come segue:
```bash
python -m transformers.onnx --model=keras-io/transformers-qa onnx/
```
Per esportare un modello memorizzato localmente, devi disporre dei pesi del modello
e file tokenizer memorizzati in una directory. Ad esempio, possiamo caricare e salvare un
checkpoint come segue:
<frameworkcontent>
<pt>
```python
>>> from transformers import AutoTokenizer, AutoModelForSequenceClassification
>>> # Load tokenizer and PyTorch weights form the Hub
>>> tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased")
>>> pt_model = AutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased")
>>> # Save to disk
>>> tokenizer.save_pretrained("local-pt-checkpoint")
>>> pt_model.save_pretrained("local-pt-checkpoint")
```
Una volta salvato il checkpoint, possiamo esportarlo su ONNX puntando l'argomento `--model`
del pacchetto `transformers.onnx` nella directory desiderata:
```bash
python -m transformers.onnx --model=local-pt-checkpoint onnx/
```
</pt>
<tf>
```python
>>> from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
>>> # Load tokenizer and TensorFlow weights from the Hub
>>> tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased")
>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased")
>>> # Save to disk
>>> tokenizer.save_pretrained("local-tf-checkpoint")
>>> tf_model.save_pretrained("local-tf-checkpoint")
```
Once the checkpoint is saved, we can export it to ONNX by pointing the `--model`
argument of the `transformers.onnx` package to the desired directory:
```bash
python -m transformers.onnx --model=local-tf-checkpoint onnx/
```
</tf>
</frameworkcontent>
### Selezione delle caratteristiche per diverse topologie di modello
Ogni configurazione giร pronta viene fornita con una serie di _caratteristiche_ che ti consentono di
esportare modelli per diversi tipi di topologie o attivitร . Come mostrato nella tabella
di seguito, ogni caratteristica รจ associata a una diversa Auto Class:
| Caratteristica | Auto Class |
| ------------------------------------ | ------------------------------------ |
| `causal-lm`, `causal-lm-with-past` | `AutoModelForCausalLM` |
| `default`, `default-with-past` | `AutoModel` |
| `masked-lm` | `AutoModelForMaskedLM` |
| `question-answering` | `AutoModelForQuestionAnswering` |
| `seq2seq-lm`, `seq2seq-lm-with-past` | `AutoModelForSeq2SeqLM` |
| `sequence-classification` | `AutoModelForSequenceClassification` |
| `token-classification` | `AutoModelForTokenClassification` |
Per ciascuna configurazione, puoi trovare l'elenco delle funzionalitร supportate tramite il
`FeaturesManager`. Ad esempio, per DistilBERT abbiamo:
```python
>>> from transformers.onnx.features import FeaturesManager
>>> distilbert_features = list(FeaturesManager.get_supported_features_for_model_type("distilbert").keys())
>>> print(distilbert_features)
["default", "masked-lm", "causal-lm", "sequence-classification", "token-classification", "question-answering"]
```
Puoi quindi passare una di queste funzionalitร all'argomento `--feature` nel
pacchetto `transformers.onnx`. Ad esempio, per esportare un modello di classificazione del testo
possiamo scegliere un modello ottimizzato dall'Hub ed eseguire:
```bash
python -m transformers.onnx --model=distilbert-base-uncased-finetuned-sst-2-english \
--feature=sequence-classification onnx/
```
che visualizzerร i seguenti registri:
```bash
Validating ONNX model...
-[โ] ONNX model output names match reference model ({'logits'})
- Validating ONNX Model output "logits":
-[โ] (2, 2) matches (2, 2)
-[โ] all values close (atol: 1e-05)
All good, model saved at: onnx/model.onnx
```
Puoi notare che in questo caso, i nomi di output del modello ottimizzato sono
`logits` invece di `last_hidden_state` che abbiamo visto con il
checkpoint `distilbert-base-uncased` precedente. Questo รจ previsto dal
modello ottimizato visto che ha una testa di e.
<Tip>
Le caratteristiche che hanno un suffisso `wtih-past` (ad es. `causal-lm-with-past`)
corrispondono a topologie di modello con stati nascosti precalcolati (chiave e valori
nei blocchi di attenzione) che possono essere utilizzati per la decodifica autoregressiva veloce.
</Tip>
### Esportazione di un modello per un'architettura non supportata
Se desideri esportare un modello la cui architettura non รจ nativamente supportata dalla
libreria, ci sono tre passaggi principali da seguire:
1. Implementare una configurazione ONNX personalizzata.
2. Esportare il modello in ONNX.
3. Convalidare gli output di PyTorch e dei modelli esportati.
In questa sezione, vedremo come DistilBERT รจ stato implementato per mostrare cosa รจ
coinvolto in ogni passaggio.
#### Implementazione di una configurazione ONNX personalizzata
Iniziamo con l'oggetto di configurazione ONNX. Forniamo tre classi
astratte da cui ereditare, a seconda del tipo di archittettura
del modello che desideri esportare:
* I modelli basati su encoder ereditano da [`~onnx.config.OnnxConfig`]
* I modelli basati su decoder ereditano da [`~onnx.config.OnnxConfigWithPast`]
* I modelli encoder-decoder ereditano da[`~onnx.config.OnnxSeq2SeqConfigWithPast`]
<Tip>
Un buon modo per implementare una configurazione ONNX personalizzata รจ guardare l'implementazione
esistente nel file `configuration_<model_name>.py` di un'architettura simile.
</Tip>
Poichรฉ DistilBERT รจ un modello basato su encoder, la sua configurazione eredita da
`OnnxConfig`:
```python
>>> from typing import Mapping, OrderedDict
>>> from transformers.onnx import OnnxConfig
>>> class DistilBertOnnxConfig(OnnxConfig):
... @property
... def inputs(self) -> Mapping[str, Mapping[int, str]]:
... return OrderedDict(
... [
... ("input_ids", {0: "batch", 1: "sequence"}),
... ("attention_mask", {0: "batch", 1: "sequence"}),
... ]
... )
```
Ogni oggetto di configurazione deve implementare la proprietร `inputs` e restituire una
mappatura, dove ogni chiave corrisponde a un input previsto e ogni valore
indica l'asse di quell'input. Per DistilBERT, possiamo vedere che sono richiesti
due input: `input_ids` e `attention_mask`. Questi inputs hanno la stessa forma di
`(batch_size, sequence_length)` per questo motivo vediamo gli stessi assi usati nella
configurazione.
<Tip>
Puoi notare che la proprietร `inputs` per `DistilBertOnnxConfig` restituisce un
`OrdinatoDict`. Ciรฒ garantisce che gli input corrispondano alla loro posizione
relativa all'interno del metodo `PreTrainedModel.forward()` durante il tracciamento del grafico.
Raccomandiamo di usare un `OrderedDict` per le proprietร `inputs` e `outputs`
quando si implementano configurazioni ONNX personalizzate.
</Tip>
Dopo aver implementato una configurazione ONNX, รจ possibile istanziarla
fornendo alla configurazione del modello base come segue:
```python
>>> from transformers import AutoConfig
>>> config = AutoConfig.from_pretrained("distilbert-base-uncased")
>>> onnx_config = DistilBertOnnxConfig(config)
```
L'oggetto risultante ha diverse proprietร utili. Ad esempio รจ possibile visualizzare il
Set operatore ONNX che verrร utilizzato durante l'esportazione:
```python
>>> print(onnx_config.default_onnx_opset)
11
```
ร inoltre possibile visualizzare gli output associati al modello come segue:
```python
>>> print(onnx_config.outputs)
OrderedDict([("last_hidden_state", {0: "batch", 1: "sequence"})])
```
Puoi notare che la proprietร degli output segue la stessa struttura degli input; esso
restituisce un `OrderedDict` di output con nome e le loro forme. La struttura di output
รจ legato alla scelta della funzione con cui viene inizializzata la configurazione.
Per impostazione predefinita, la configurazione ONNX viene inizializzata con la funzione 'predefinita'
che corrisponde all'esportazione di un modello caricato con la classe `AutoModel`. Se tu
desideri esportare una topologia di modello diversa, รจ sufficiente fornire una funzionalitร diversa a
l'argomento `task` quando inizializzi la configurazione ONNX. Ad esempio, se
volevamo esportare DistilBERT con una testa di classificazione per sequenze, potremmo
usare:
```python
>>> from transformers import AutoConfig
>>> config = AutoConfig.from_pretrained("distilbert-base-uncased")
>>> onnx_config_for_seq_clf = DistilBertOnnxConfig(config, task="sequence-classification")
>>> print(onnx_config_for_seq_clf.outputs)
OrderedDict([('logits', {0: 'batch'})])
```
<Tip>
Tutte le proprietร e i metodi di base associati a [`~onnx.config.OnnxConfig`] e le
altre classi di configurazione possono essere sovrascritte se necessario. Guarda
[`BartOnnxConfig`] per un esempio avanzato.
</Tip>
#### Esportazione del modello
Una volta implementata la configurazione ONNX, il passaggio successivo consiste nell'esportare il
modello. Qui possiamo usare la funzione `export()` fornita dal
pacchetto `transformers.onnx`. Questa funzione prevede la configurazione ONNX, insieme
con il modello base e il tokenizer e il percorso per salvare il file esportato:
```python
>>> from pathlib import Path
>>> from transformers.onnx import export
>>> from transformers import AutoTokenizer, AutoModel
>>> onnx_path = Path("model.onnx")
>>> model_ckpt = "distilbert-base-uncased"
>>> base_model = AutoModel.from_pretrained(model_ckpt)
>>> tokenizer = AutoTokenizer.from_pretrained(model_ckpt)
>>> onnx_inputs, onnx_outputs = export(tokenizer, base_model, onnx_config, onnx_config.default_onnx_opset, onnx_path)
```
Gli `onnx_inputs` e `onnx_outputs` restituiti dalla funzione `export()` sono
liste di chiavi definite nelle proprietร di `input` e `output` della
configurazione. Una volta esportato il modello, puoi verificare che il modello sia ben
formato come segue:
```python
>>> import onnx
>>> onnx_model = onnx.load("model.onnx")
>>> onnx.checker.check_model(onnx_model)
```
<Tip>
Se il tuo modello รจ piรน largo di 2 GB, vedrai che molti file aggiuntivi sono
creati durante l'esportazione. Questo รจ _previsto_ perchรฉ ONNX utilizza [Protocol
Buffer](https://developers.google.com/protocol-buffers/) per memorizzare il modello e
questi hanno un limite di dimensione 2 GB. Vedi la [Documentazione
ONNX](https://github.com/onnx/onnx/blob/master/docs/ExternalData.md)
per istruzioni su come caricare modelli con dati esterni.
</Tip>
#### Convalida degli output del modello
Il passaggio finale consiste nel convalidare gli output dal modello di base e quello esportato
corrispondere entro una soglia di tolleranza assoluta. Qui possiamo usare la
Funzione `validate_model_outputs()` fornita dal pacchetto `transformers.onnx`
come segue:
```python
>>> from transformers.onnx import validate_model_outputs
>>> validate_model_outputs(
... onnx_config, tokenizer, base_model, onnx_path, onnx_outputs, onnx_config.atol_for_validation
... )
```
Questa funzione usa il metodo `OnnxConfig.generate_dummy_inputs()` per generare
input per il modello di base e quello esportato e la tolleranza assoluta puรฒ essere
definita nella configurazione. Generalmente troviamo una corrispondenza numerica nell'intervallo da 1e-6
a 1e-4, anche se รจ probabile che qualsiasi cosa inferiore a 1e-3 vada bene.
### Contribuire con una nuova configurazione a ๐ค Transformers
Stiamo cercando di espandere l'insieme di configurazioni giร pronte e di accettare
contributi della community! Se vuoi contribuire con la tua aggiunta
nella libreria, dovrai:
* Implementare la configurazione ONNX nella corrispondente `configuration file
_<model_name>.py`
* Includere l'architettura del modello e le funzioni corrispondenti in [`~onnx.features.FeatureManager`]
* Aggiungere la tua architettura del modello ai test in `test_onnx_v2.py`
Scopri come stato contribuito la configurazione per [IBERT]
(https://github.com/huggingface/transformers/pull/14868/files) per
avere un'idea di cosa รจ coinvolto.
## TorchScript
<Tip>
Questo รจ l'inizio dei nostri esperimenti con TorchScript e stiamo ancora esplorando le sue capacitร con
modelli con variable-input-size. ร una nostra prioritร e approfondiremo le nostre analisi nelle prossime versioni,
con piรน esempi di codici, un'implementazione piรน flessibile e benchmark che confrontano i codici basati su Python con quelli compilati con
TorchScript.
</Tip>
Secondo la documentazione di Pytorch: "TorchScript รจ un modo per creare modelli serializzabili e ottimizzabili da codice
Pytorch". I due moduli di Pytorch [JIT e TRACE](https://pytorch.org/docs/stable/jit.html) consentono allo sviluppatore di esportare
il loro modello da riutilizzare in altri programmi, come i programmi C++ orientati all'efficienza.
Abbiamo fornito un'interfaccia che consente l'esportazione di modelli ๐ค Transformers in TorchScript in modo che possano essere riutilizzati
in un ambiente diverso rispetto a un programma Python basato su Pytorch. Qui spieghiamo come esportare e utilizzare i nostri modelli utilizzando
TorchScript.
Esportare un modello richiede due cose:
- Un passaggio in avanti con input fittizzi.
- Istanziazione del modello con flag `torchscript`.
Queste necessitร implicano diverse cose a cui gli sviluppatori dovrebbero prestare attenzione. Questi dettagli mostrati sotto.
### Flag TorchScript e pesi legati
Questo flag รจ necessario perchรฉ la maggior parte dei modelli linguistici in questo repository hanno pesi legati tra il loro
strato "Embedding" e lo strato "Decoding". TorchScript non consente l'esportazione di modelli che hanno pesi
legati, quindi รจ necessario prima slegare e clonare i pesi.
Ciรฒ implica che i modelli istanziati con il flag `torchscript` hanno il loro strato `Embedding` e strato `Decoding`
separato, il che significa che non dovrebbero essere addestrati in futuro. L'allenamento de-sincronizza i due
strati, portando a risultati inaspettati.
Questo non รจ il caso per i modelli che non hanno una testa del modello linguistico, poichรฉ quelli non hanno pesi legati. Questi modelli
puรฒ essere esportato in sicurezza senza il flag `torchscript`.
### Input fittizi e standard lengths
Gli input fittizzi sono usati per fare un modello passaggio in avanti . Mentre i valori degli input si propagano attraverso i strati,
Pytorch tiene traccia delle diverse operazioni eseguite su ciascun tensore. Queste operazioni registrate vengono quindi utilizzate per
creare la "traccia" del modello.
La traccia viene creata relativamente alle dimensioni degli input. ร quindi vincolato dalle dimensioni dell'input
fittizio e non funzionerร per altre lunghezze di sequenza o dimensioni batch. Quando si proverร con una dimensione diversa, ci sarร errore
come:
`La dimensione espansa del tensore (3) deve corrispondere alla dimensione esistente (7) nella dimensione non singleton 2`
will be raised. Si consiglia pertanto di tracciare il modello con una dimensione di input fittizia grande almeno quanto il piรน grande
input che verrร fornito al modello durante l'inferenza. ร possibile eseguire il padding per riempire i valori mancanti. Il modello
sarร tracciato con una grande dimensione di input, tuttavia, anche le dimensioni della diverse matrici saranno grandi,
risultando in piรน calcoli.
Si raccomanda di prestare attenzione al numero totale di operazioni eseguite su ciascun input e di seguire da vicino le prestazioni
durante l'esportazione di modelli di sequenza-lunghezza variabili.
### Usare TorchSscript in Python
Di seguito รจ riportato un esempio, che mostra come salvare, caricare modelli e come utilizzare la traccia per l'inferenza.
#### Salvare un modello
Questo frammento di codice mostra come usare TorchScript per esportare un `BertModel`. Qui il `BertModel` รจ istanziato secondo
una classe `BertConfig` e quindi salvato su disco con il nome del file `traced_bert.pt`
```python
from transformers import BertModel, BertTokenizer, BertConfig
import torch
enc = BertTokenizer.from_pretrained("bert-base-uncased")
# Tokenizing input text
text = "[CLS] Who was Jim Henson ? [SEP] Jim Henson was a puppeteer [SEP]"
tokenized_text = enc.tokenize(text)
# Masking one of the input tokens
masked_index = 8
tokenized_text[masked_index] = "[MASK]"
indexed_tokens = enc.convert_tokens_to_ids(tokenized_text)
segments_ids = [0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1]
# Creating a dummy input
tokens_tensor = torch.tensor([indexed_tokens])
segments_tensors = torch.tensor([segments_ids])
dummy_input = [tokens_tensor, segments_tensors]
# Initializing the model with the torchscript flag
# Flag set to True even though it is not necessary as this model does not have an LM Head.
config = BertConfig(
vocab_size_or_config_json_file=32000,
hidden_size=768,
num_hidden_layers=12,
num_attention_heads=12,
intermediate_size=3072,
torchscript=True,
)
# Instantiating the model
model = BertModel(config)
# The model needs to be in evaluation mode
model.eval()
# If you are instantiating the model with *from_pretrained* you can also easily set the TorchScript flag
model = BertModel.from_pretrained("bert-base-uncased", torchscript=True)
# Creating the trace
traced_model = torch.jit.trace(model, [tokens_tensor, segments_tensors])
torch.jit.save(traced_model, "traced_bert.pt")
```
#### Caricare un modello
Questo frammento di codice mostra come caricare il `BertModel` che era stato precedentemente salvato su disco con il nome `traced_bert.pt`.
Stiamo riutilizzando il `dummy_input` precedentemente inizializzato.
```python
loaded_model = torch.jit.load("traced_bert.pt")
loaded_model.eval()
all_encoder_layers, pooled_output = loaded_model(*dummy_input)
```
#### Utilizzare un modello tracciato per l'inferenza
Usare il modello tracciato per l'inferenza รจ semplice come usare il suo metodo dunder `__call__`:
```python
traced_model(tokens_tensor, segments_tensors)
```
###Implementare modelli HuggingFace TorchScript su AWS utilizzando Neuron SDK
AWS ha introdotto [Amazon EC2 Inf1](https://aws.amazon.com/ec2/instance-types/inf1/)
famiglia di istanze per l'inferenza di machine learning a basso costo e ad alte prestazioni nel cloud.
Le istanze Inf1 sono alimentate dal chip AWS Inferentia, un acceleratore hardware personalizzato,
specializzato in carichi di lavoro di inferenza di deep learning.
[AWS Neuron](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/#)
รจ l'SDK per Inferentia che supporta il tracciamento e l'ottimizzazione dei modelli transformers per
distribuzione su Inf1. L'SDK Neuron fornisce:
1. API di facile utilizzo con una riga di modifica del codice per tracciare e ottimizzare un modello TorchScript per l'inferenza nel cloud.
2. Ottimizzazioni delle prestazioni pronte all'uso per [miglioramento dei costi-prestazioni](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/neuron-guide/benchmark/>)
3. Supporto per i modelli di trasformatori HuggingFace costruiti con [PyTorch](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/src/examples/pytorch/bert_tutorial/tutorial_pretrained_bert.html)
o [TensorFlow](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/src/examples/tensorflow/huggingface_bert/huggingface_bert.html).
#### Implicazioni
Modelli Transformers basati su architettura [BERT (Bidirectional Encoder Representations from Transformers)](https://huggingface.co/docs/transformers/main/model_doc/bert),
o sue varianti come [distilBERT](https://huggingface.co/docs/transformers/main/model_doc/distilbert)
e [roBERTa](https://huggingface.co/docs/transformers/main/model_doc/roberta)
funzioneranno meglio su Inf1 per attivitร non generative come la question answering estrattive,
Classificazione della sequenza, Classificazione dei token. In alternativa, generazione di testo
le attivitร possono essere adattate per essere eseguite su Inf1, secondo questo [tutorial AWS Neuron MarianMT](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/src/examples/pytorch/transformers-marianmt.html).
Ulteriori informazioni sui modelli che possono essere convertiti fuori dagli schemi su Inferentia possono essere
trovati nella [sezione Model Architecture Fit della documentazione Neuron](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/neuron-guide/models/models-inferentia.html#models-inferentia).
#### Dipendenze
L'utilizzo di AWS Neuron per convertire i modelli richiede le seguenti dipendenze e l'ambiente:
* A [Neuron SDK environment](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/neuron-guide/neuron-frameworks/pytorch-neuron/index.html#installation-guide),
which comes pre-configured on [AWS Deep Learning AMI](https://docs.aws.amazon.com/dlami/latest/devguide/tutorial-inferentia-launching.html).
#### Convertire un modello per AWS Neuron
Usando lo stesso script come in [Usando TorchScipt in Python](https://huggingface.co/docs/transformers/main/en/serialization#using-torchscript-in-python)
per tracciare un "BertModel", importi l'estensione del framework `torch.neuron` per accedere
i componenti di Neuron SDK tramite un'API Python.
```python
from transformers import BertModel, BertTokenizer, BertConfig
import torch
import torch.neuron
```
E modificare solo la riga di codice di traccia
Da:
```python
torch.jit.trace(model, [tokens_tensor, segments_tensors])
```
A:
```python
torch.neuron.trace(model, [token_tensor, segments_tensors])
```
Questa modifica consente a Neuron SDK di tracciare il modello e ottimizzarlo per l'esecuzione nelle istanze Inf1.
Per ulteriori informazioni sulle funzionalitร , gli strumenti, i tutorial di esempi e gli ultimi aggiornamenti di AWS Neuron SDK,
consultare la [documentazione AWS NeuronSDK](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/index.html).
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/converting_tensorflow_models.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Convertire checkpoint di Tensorflow
ร disponibile un'interfaccia a linea di comando per convertire gli originali checkpoint di Bert/GPT/GPT-2/Transformer-XL/XLNet/XLM
in modelli che possono essere caricati utilizzando i metodi `from_pretrained` della libreria.
<Tip>
A partire dalla versione 2.3.0 lo script di conversione รจ parte di transformers CLI (**transformers-cli**), disponibile in ogni installazione
di transformers >=2.3.0.
La seguente documentazione riflette il formato dei comandi di **transformers-cli convert**.
</Tip>
## BERT
Puoi convertire qualunque checkpoint Tensorflow di BERT (in particolare
[i modeli pre-allenati rilasciati da Google](https://github.com/google-research/bert#pre-trained-models))
in un file di salvataggio Pytorch utilizzando lo script
[convert_bert_original_tf_checkpoint_to_pytorch.py](https://github.com/huggingface/transformers/tree/main/src/transformers/models/bert/convert_bert_original_tf_checkpoint_to_pytorch.py).
Questo CLI prende come input un checkpoint di Tensorflow (tre files che iniziano con `bert_model.ckpt`) ed il relativo
file di configurazione (`bert_config.json`), crea un modello Pytorch per questa configurazione, carica i pesi dal
checkpoint di Tensorflow nel modello di Pytorch e salva il modello che ne risulta in un file di salvataggio standard di Pytorch che
puรฒ essere importato utilizzando `from_pretrained()` (vedi l'esempio nel
[quicktour](quicktour) , [run_glue.py](https://github.com/huggingface/transformers/tree/main/examples/pytorch/text-classification/run_glue.py) ).
Devi soltanto lanciare questo script di conversione **una volta** per ottenere un modello Pytorch. Dopodichรจ, potrai tralasciare
il checkpoint di Tensorflow (i tre files che iniziano con `bert_model.ckpt`), ma assicurati di tenere il file di configurazione
(`bert_config.json`) ed il file di vocabolario (`vocab.txt`) in quanto queste componenti sono necessarie anche per il modello di Pytorch.
Per lanciare questo specifico script di conversione avrai bisogno di un'installazione di Tensorflow e di Pytorch
(`pip install tensorflow`). Il resto della repository richiede soltanto Pytorch.
Questo รจ un esempio del processo di conversione per un modello `BERT-Base Uncased` pre-allenato:
```bash
export BERT_BASE_DIR=/path/to/bert/uncased_L-12_H-768_A-12
transformers-cli convert --model_type bert \
--tf_checkpoint $BERT_BASE_DIR/bert_model.ckpt \
--config $BERT_BASE_DIR/bert_config.json \
--pytorch_dump_output $BERT_BASE_DIR/pytorch_model.bin
```
Puoi scaricare i modelli pre-allenati di Google per la conversione [qua](https://github.com/google-research/bert#pre-trained-models).
## ALBERT
Per il modello ALBERT, converti checkpoint di Tensoflow in Pytorch utilizzando lo script
[convert_albert_original_tf_checkpoint_to_pytorch.py](https://github.com/huggingface/transformers/tree/main/src/transformers/models/albert/convert_albert_original_tf_checkpoint_to_pytorch.py).
Il CLI prende come input un checkpoint di Tensorflow (tre files che iniziano con `model.ckpt-best`) e i relativi file di
configurazione (`albert_config.json`), dopodichรจ crea e salva un modello Pytorch. Per lanciare questa conversione
avrai bisogno di un'installazione di Tensorflow e di Pytorch.
Ecco un esempio del procedimento di conversione di un modello `ALBERT Base` pre-allenato:
```bash
export ALBERT_BASE_DIR=/path/to/albert/albert_base
transformers-cli convert --model_type albert \
--tf_checkpoint $ALBERT_BASE_DIR/model.ckpt-best \
--config $ALBERT_BASE_DIR/albert_config.json \
--pytorch_dump_output $ALBERT_BASE_DIR/pytorch_model.bin
```
Puoi scaricare i modelli pre-allenati di Google per la conversione [qui](https://github.com/google-research/albert#pre-trained-models).
## OpenAI GPT
Ecco un esempio del processo di conversione di un modello OpenAI GPT pre-allenato, assumendo che il tuo checkpoint di NumPy
sia salvato nello stesso formato dei modelli pre-allenati OpenAI (vedi [qui](https://github.com/openai/finetune-transformer-lm)):
```bash
export OPENAI_GPT_CHECKPOINT_FOLDER_PATH=/path/to/openai/pretrained/numpy/weights
transformers-cli convert --model_type gpt \
--tf_checkpoint $OPENAI_GPT_CHECKPOINT_FOLDER_PATH \
--pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
[--config OPENAI_GPT_CONFIG] \
[--finetuning_task_name OPENAI_GPT_FINETUNED_TASK] \
```
## OpenAI GPT-2
Ecco un esempio del processo di conversione di un modello OpenAI GPT-2 pre-allenato (vedi [qui](https://github.com/openai/gpt-2)):
```bash
export OPENAI_GPT2_CHECKPOINT_PATH=/path/to/gpt2/pretrained/weights
transformers-cli convert --model_type gpt2 \
--tf_checkpoint $OPENAI_GPT2_CHECKPOINT_PATH \
--pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
[--config OPENAI_GPT2_CONFIG] \
[--finetuning_task_name OPENAI_GPT2_FINETUNED_TASK]
```
## XLNet
Ecco un esempio del processo di conversione di un modello XLNet pre-allenato:
```bash
export TRANSFO_XL_CHECKPOINT_PATH=/path/to/xlnet/checkpoint
export TRANSFO_XL_CONFIG_PATH=/path/to/xlnet/config
transformers-cli convert --model_type xlnet \
--tf_checkpoint $TRANSFO_XL_CHECKPOINT_PATH \
--config $TRANSFO_XL_CONFIG_PATH \
--pytorch_dump_output $PYTORCH_DUMP_OUTPUT \
[--finetuning_task_name XLNET_FINETUNED_TASK] \
```
## XLM
Ecco un esempio del processo di conversione di un modello XLM pre-allenato:
```bash
export XLM_CHECKPOINT_PATH=/path/to/xlm/checkpoint
transformers-cli convert --model_type xlm \
--tf_checkpoint $XLM_CHECKPOINT_PATH \
--pytorch_dump_output $PYTORCH_DUMP_OUTPUT
[--config XML_CONFIG] \
[--finetuning_task_name XML_FINETUNED_TASK]
```
## T5
Ecco un esempio del processo di conversione di un modello T5 pre-allenato:
```bash
export T5=/path/to/t5/uncased_L-12_H-768_A-12
transformers-cli convert --model_type t5 \
--tf_checkpoint $T5/t5_model.ckpt \
--config $T5/t5_config.json \
--pytorch_dump_output $T5/pytorch_model.bin
```
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/autoclass_tutorial.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Carica istanze pre-allenate con AutoClass
Con cosรฌ tante architetture Transformer differenti, puรฒ essere sfidante crearne una per il tuo checkpoint. Come parte della filosofia centrale di ๐ค Transformers per rendere la libreria facile, semplice e flessibile da utilizzare, una `AutoClass` inferisce e carica automaticamente l'architettura corretta da un dato checkpoint. Il metodo `from_pretrained` ti permette di caricare velocemente un modello pre-allenato per qualsiasi architettura, cosรฌ non devi utilizzare tempo e risorse per allenare un modello da zero. Produrre questo codice agnostico ai checkpoint significa che se il tuo codice funziona per un checkpoint, funzionerร anche per un altro checkpoint, purchรฉ sia stato allenato per un compito simile, anche se l'architettura รจ differente.
<Tip>
Ricorda, con architettura ci si riferisce allo scheletro del modello e con checkpoint ai pesi di una determinata architettura. Per esempio, [BERT](https://huggingface.co/bert-base-uncased) รจ un'architettura, mentre `bert-base-uncased` รจ un checkpoint. Modello รจ un termine generale che puรฒ significare sia architettura che checkpoint.
</Tip>
In questo tutorial, imparerai a:
* Caricare un tokenizer pre-allenato.
* Caricare un estrattore di caratteristiche (feature extractor, in inglese) pre-allenato.
* Caricare un processore pre-allenato.
* Caricare un modello pre-allenato.
## AutoTokenizer
Quasi tutti i compiti di NLP iniziano con un tokenizer. Un tokenizer converte il tuo input in un formato che possa essere elaborato dal modello.
Carica un tokenizer con [`AutoTokenizer.from_pretrained`]:
```py
>>> from transformers import AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("xlm-roberta-base")
```
Poi tokenizza il tuo input come mostrato in seguito:
```py
>>> sequenza = "In un buco nel terreno viveva uno Hobbit."
>>> print(tokenizer(sequenza))
{'input_ids': [0, 360, 51, 373, 587, 1718, 54644, 22597, 330, 3269, 2291, 22155, 18, 5, 2],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}
```
## AutoFeatureExtractor
Per compiti inerenti a audio e video, un feature extractor processa il segnale audio o l'immagine nel formato di input corretto.
Carica un feature extractor con [`AutoFeatureExtractor.from_pretrained`]:
```py
>>> from transformers import AutoFeatureExtractor
>>> feature_extractor = AutoFeatureExtractor.from_pretrained(
... "ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition"
... )
```
## AutoProcessor
Compiti multimodali richiedono un processore che combini i due tipi di strumenti di elaborazione. Per esempio, il modello [LayoutLMV2](model_doc/layoutlmv2) richiede un feature extractor per gestire le immagine e un tokenizer per gestire il testo; un processore li combina entrambi.
Carica un processore con [`AutoProcessor.from_pretrained`]:
```py
>>> from transformers import AutoProcessor
>>> processor = AutoProcessor.from_pretrained("microsoft/layoutlmv2-base-uncased")
```
## AutoModel
<frameworkcontent>
<pt>
Infine, le classi `AutoModelFor` ti permettono di caricare un modello pre-allenato per un determinato compito (guarda [qui](model_doc/auto) per una lista completa di compiti presenti). Per esempio, carica un modello per la classificazione di sequenze con [`AutoModelForSequenceClassification.from_pretrained`]:
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased")
```
Semplicemente utilizza lo stesso checkpoint per caricare un'architettura per un task differente:
```py
>>> from transformers import AutoModelForTokenClassification
>>> model = AutoModelForTokenClassification.from_pretrained("distilbert-base-uncased")
```
Generalmente, raccomandiamo di utilizzare la classe `AutoTokenizer` e la classe `AutoModelFor` per caricare istanze pre-allenate dei modelli. Questo ti assicurerร di aver caricato la corretta architettura ogni volta. Nel prossimo [tutorial](preprocessing), imparerai come utilizzare il tokenizer, il feature extractor e il processore per elaborare un dataset per il fine-tuning.
</pt>
<tf>
Infine, le classi `TFAutoModelFor` ti permettono di caricare un modello pre-allenato per un determinato compito (guarda [qui](model_doc/auto) per una lista completa di compiti presenti). Per esempio, carica un modello per la classificazione di sequenze con [`TFAutoModelForSequenceClassification.from_pretrained`]:
```py
>>> from transformers import TFAutoModelForSequenceClassification
>>> model = TFAutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased")
```
Semplicemente utilizza lo stesso checkpoint per caricare un'architettura per un task differente:
```py
>>> from transformers import TFAutoModelForTokenClassification
>>> model = TFAutoModelForTokenClassification.from_pretrained("distilbert-base-uncased")
```
Generalmente, raccomandiamo di utilizzare la classe `AutoTokenizer` e la classe `TFAutoModelFor` per caricare istanze pre-allenate dei modelli. Questo ti assicurerร di aver caricato la corretta architettura ogni volta. Nel prossimo [tutorial](preprocessing), imparerai come utilizzare il tokenizer, il feature extractor e il processore per elaborare un dataset per il fine-tuning.
</tf>
</frameworkcontent>
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/perf_infer_cpu.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Inferenza Efficiente su CPU
Questa guida si concentra sull'inferenza di modelli di grandi dimensioni in modo efficiente sulla CPU.
## `BetterTransformer` per inferenza piรน rapida
Abbiamo integrato di recente `BetterTransformer` per fare inferenza piรน rapidamente con modelli per testi, immagini e audio. Visualizza la documentazione sull'integrazione [qui](https://huggingface.co/docs/optimum/bettertransformer/overview) per maggiori dettagli.
## PyTorch JIT-mode (TorchScript)
TorchScript รจ un modo di creare modelli serializzabili e ottimizzabili da codice PyTorch. Ogni programmma TorchScript puรฒ esere salvato da un processo Python e caricato in un processo dove non ci sono dipendenze Python.
Comparandolo con l'eager mode di default, jit mode in PyTorch normalmente fornisce prestazioni migliori per l'inferenza del modello da parte di metodologie di ottimizzazione come la operator fusion.
Per una prima introduzione a TorchScript, vedi la Introduction to [PyTorch TorchScript tutorial](https://pytorch.org/tutorials/beginner/Intro_to_TorchScript_tutorial.html#tracing-modules).
### IPEX Graph Optimization con JIT-mode
Intelยฎ Extension per PyTorch fornnisce ulteriori ottimizzazioni in jit mode per i modelli della serie Transformers. Consigliamo vivamente agli utenti di usufruire dei vantaggi di Intelยฎ Extension per PyTorch con jit mode. Alcuni operator patterns usati fequentemente dai modelli Transformers models sono giร supportati in Intelยฎ Extension per PyTorch con jit mode fusions. Questi fusion patterns come Multi-head-attention fusion, Concat Linear, Linear+Add, Linear+Gelu, Add+LayerNorm fusion and etc. sono abilitati e hanno buone performance. I benefici della fusion รจ fornito agli utenti in modo trasparente. In base alle analisi, il ~70% dei problemi piรน popolari in NLP question-answering, text-classification, and token-classification possono avere benefici sulle performance grazie ai fusion patterns sia per Float32 precision che per BFloat16 Mixed precision.
Vedi maggiori informazioni per [IPEX Graph Optimization](https://intel.github.io/intel-extension-for-pytorch/cpu/latest/tutorials/features/graph_optimization.html).
#### Installazione di IPEX
I rilasci di IPEX seguono PyTorch, verifica i vari approcci per [IPEX installation](https://intel.github.io/intel-extension-for-pytorch/).
### Utilizzo del JIT-mode
Per abilitare JIT-mode in Trainer per evaluation e prediction, devi aggiungere `jit_mode_eval` negli argomenti di Trainer.
<Tip warning={true}>
per PyTorch >= 1.14.0. JIT-mode potrebe giovare a qualsiasi modello di prediction e evaluaion visto che il dict input รจ supportato in jit.trace
per PyTorch < 1.14.0. JIT-mode potrebbe giovare ai modelli il cui ordine dei parametri corrisponde all'ordine delle tuple in ingresso in jit.trace, come i modelli per question-answering.
Nel caso in cui l'ordine dei parametri seguenti non corrisponda all'ordine delle tuple in ingresso in jit.trace, come nei modelli di text-classification, jit.trace fallirร e lo cattureremo con una eccezione al fine di renderlo un fallback. Il logging รจ usato per notificare gli utenti.
</Tip>
Trovi un esempo con caso d'uso in [Transformers question-answering](https://github.com/huggingface/transformers/tree/main/examples/pytorch/question-answering)
- Inference using jit mode on CPU:
<pre>python run_qa.py \
--model_name_or_path csarron/bert-base-uncased-squad-v1 \
--dataset_name squad \
--do_eval \
--max_seq_length 384 \
--doc_stride 128 \
--output_dir /tmp/ \
--no_cuda \
<b>--jit_mode_eval </b></pre>
- Inference with IPEX using jit mode on CPU:
<pre>python run_qa.py \
--model_name_or_path csarron/bert-base-uncased-squad-v1 \
--dataset_name squad \
--do_eval \
--max_seq_length 384 \
--doc_stride 128 \
--output_dir /tmp/ \
--no_cuda \
<b>--use_ipex \</b>
<b>--jit_mode_eval</b></pre>
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/accelerate.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Allenamento distribuito con ๐ค Accelerate
La parallelizzazione รจ emersa come strategia per allenare modelli sempre piรน grandi su hardware limitato e accelerarne la velocitร di allenamento di diversi ordini di magnitudine. In Hugging Face, abbiamo creato la libreria [๐ค Accelerate](https://huggingface.co/docs/accelerate) per aiutarti ad allenare in modo semplice un modello ๐ค Transformers su qualsiasi tipo di configurazione distribuita, sia che si tratti di piรน GPU su una sola macchina o di piรน GPU su piรน macchine. In questo tutorial, imparerai come personalizzare il training loop nativo di PyTorch per consentire l'addestramento in un ambiente distribuito.
## Configurazione
Inizia installando ๐ค Accelerate:
```bash
pip install accelerate
```
Poi importa e crea un oggetto [`Accelerator`](https://huggingface.co/docs/accelerate/package_reference/accelerator#accelerate.Accelerator). `Accelerator` rileverร automaticamente il tuo setup distribuito e inizializzerร tutte le componenti necessarie per l'allenamento. Non dovrai allocare esplicitamente il tuo modello su un device.
```py
>>> from accelerate import Accelerator
>>> accelerator = Accelerator()
```
## Preparati ad accelerare
Il prossimo passo รจ quello di passare tutti gli oggetti rilevanti per l'allenamento al metodo [`prepare`](https://huggingface.co/docs/accelerate/package_reference/accelerator#accelerate.Accelerator.prepare). Questo include i tuoi DataLoaders per l'allenamento e per la valutazione, un modello e un ottimizzatore:
```py
>>> train_dataloader, eval_dataloader, model, optimizer = accelerator.prepare(
... train_dataloader, eval_dataloader, model, optimizer
... )
```
## Backward
Infine, sostituisci il tipico metodo `loss.backward()` nel tuo loop di allenamento con il metodo [`backward`](https://huggingface.co/docs/accelerate/package_reference/accelerator#accelerate.Accelerator.backward) di ๐ค Accelerate:
```py
>>> for epoch in range(num_epochs):
... for batch in train_dataloader:
... outputs = model(**batch)
... loss = outputs.loss
... accelerator.backward(loss)
... optimizer.step()
... lr_scheduler.step()
... optimizer.zero_grad()
... progress_bar.update(1)
```
Come puoi vedere nel seguente codice, hai solo bisogno di aggiungere quattro righe in piรน di codice al tuo training loop per abilitare l'allenamento distribuito!
```diff
+ from accelerate import Accelerator
from transformers import AdamW, AutoModelForSequenceClassification, get_scheduler
+ accelerator = Accelerator()
model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2)
optimizer = AdamW(model.parameters(), lr=3e-5)
- device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
- model.to(device)
+ train_dataloader, eval_dataloader, model, optimizer = accelerator.prepare(
+ train_dataloader, eval_dataloader, model, optimizer
+ )
num_epochs = 3
num_training_steps = num_epochs * len(train_dataloader)
lr_scheduler = get_scheduler(
"linear",
optimizer=optimizer,
num_warmup_steps=0,
num_training_steps=num_training_steps
)
progress_bar = tqdm(range(num_training_steps))
model.train()
for epoch in range(num_epochs):
for batch in train_dataloader:
- batch = {k: v.to(device) for k, v in batch.items()}
outputs = model(**batch)
loss = outputs.loss
- loss.backward()
+ accelerator.backward(loss)
optimizer.step()
lr_scheduler.step()
optimizer.zero_grad()
progress_bar.update(1)
```
## Allenamento
Una volta che hai aggiunto le righe di codice rilevanti, lancia il tuo allenamento in uno script o in un notebook come Colaboratory.
### Allenamento con uno script
Se stai eseguendo il tuo allenamento da uno script, esegui il comando seguente per creare e salvare un file di configurazione:
```bash
accelerate config
```
Poi lancia il tuo allenamento con:
```bash
accelerate launch train.py
```
### Allenamento con un notebook
La libreria ๐ค Accelerate puรฒ anche essere utilizzata in un notebook se stai pianificando di utilizzare le TPU di Colaboratory. Inserisci tutto il codice legato all'allenamento in una funzione, e passala al `notebook_launcher`:
```py
>>> from accelerate import notebook_launcher
>>> notebook_launcher(training_function)
```
Per maggiori informazioni relative a ๐ค Accelerate e le sue numerose funzionalitร , fai riferimento alla [documentazione](https://huggingface.co/docs/accelerate).
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/perf_train_special.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Addestramento su Hardware Specializzato
<Tip>
Nota: Molte delle strategie introdotte nella [sezione sulla GPU singola](perf_train_gpu_one) (come mixed precision training o gradient accumulation) e [sezione multi-GPU](perf_train_gpu_many) sono generiche e applicabili all'addestramento di modelli in generale quindi assicurati di dargli un'occhiata prima di immergerti in questa sezione.
</Tip>
Questo documento sarร presto completato con informazioni su come effettuare la formazione su hardware specializzato.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/pr_checks.md
|
<!---
Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Controlli su una Pull Request
Quando apri una pull request sui ๐ค Transformers, vengono eseguiti un discreto numero di controlli per assicurarsi che la patch che stai aggiungendo non stia rompendo qualcosa di esistente. Questi controlli sono di quattro tipi:
- test regolari
- costruzione della documentazione
- stile del codice e della documentazione
- coerenza generale del repository
In questo documento, cercheremo di spiegare quali sono i vari controlli e le loro ragioni, oltre a spiegare come eseguire il debug locale se uno di essi fallisce sulla tua PR.
Nota che tutti richiedono un'installazione dev:
```bash
pip install transformers[dev]
```
o un'installazione modificabile:
```bash
pip install -e .[dev]
```
all'interno del repo Transformers.
## Tests
Tutti i job che iniziano con `ci/circleci: run_tests_` eseguono parti della suite di test dei Transformers. Ognuno di questi job si concentra su una parte della libreria in un determinato ambiente: per esempio `ci/circleci: run_tests_pipelines_tf` esegue il test delle pipeline in un ambiente in cui รจ installato solo TensorFlow.
Nota che per evitare di eseguire i test quando non ci sono cambiamenti reali nei moduli che si stanno testando, ogni volta viene eseguita solo una parte della suite di test: viene eseguita una utility per determinare le differenze nella libreria tra prima e dopo la PR (ciรฒ che GitHub mostra nella scheda "Files changes") e sceglie i test che sono stati impattati dalla diff. Questa utility puรฒ essere eseguita localmente con:
```bash
python utils/tests_fetcher.py
```
dalla root del repo Transformers. Di seguito ciรฒ che farร :
1. Controlla per ogni file nel diff se le modifiche sono nel codice o solo nei commenti o nelle docstrings. Vengono mantenuti solo i file con modifiche reali al codice.
2. Costruisce una mappa interna che fornisce per ogni file del codice sorgente della libreria tutti i file su cui ha un impatto ricorsivo. Si dice che il modulo A ha un impatto sul modulo B se il modulo B importa il modulo A. Per l'impatto ricorsivo, abbiamo bisogno di una catena di moduli che va dal modulo A al modulo B in cui ogni modulo importa il precedente.
3. Applica questa mappa ai file raccolti nel passaggio 1, si ottiene l'elenco dei file del modello interessati dalla PR.
4. Mappa ciascuno di questi file con i corrispondenti file di test e ottiene l'elenco dei test da eseguire.
Quando esegui lo script in locale, dovresti ottenere la stampa dei risultati dei passi 1, 3 e 4 e quindi sapere quali test sono stati eseguiti. Lo script creerร anche un file chiamato `test_list.txt` che contiene l'elenco dei test da eseguire e che puoi eseguire localmente con il seguente comando:
```bash
python -m pytest -n 8 --dist=loadfile -rA -s $(cat test_list.txt)
```
Nel caso in cui qualcosa sia sfuggito, l'intera suite di test viene eseguita quotidianamente.
## Build della documentazione
Il job `ci/circleci: build_doc` esegue una build della documentazione per assicurarsi che tutto sia a posto una volta che la PR รจ stata unita. Se questo passaggio fallisce, puoi controllare localmente entrando nella cartella `docs` del repo Transformers e digitare
```bash
make html
```
Sphinx non รจ noto per i suoi messaggi di errore chiari, quindi potrebbe essere necessario che provi alcune cose per trovare davvero la fonte dell'errore.
## Stile del codice e della documentazione
La formattazione del codice viene applicata a tutti i file sorgenti, agli esempi e ai test usando `black` e `isort`. Abbiamo anche uno strumento personalizzato che si occupa della formattazione delle docstring e dei file `rst` (`utils/style_doc.py`), cosรฌ come dell'ordine dei lazy imports eseguiti nei file `__init__.py` dei Transformers (`utils/custom_init_isort.py`). Tutto questo puรฒ essere lanciato eseguendo
```bash
make style
```
I controlli della CI sono applicati all'interno del controllo `ci/circleci: check_code_quality`. Esegue anche `flake8`, che dร un'occhiata di base al codice e si lamenta se trova una variabile non definita o non utilizzata. Per eseguire questo controllo localmente, usare
```bash
make quality
```
Questa operazione puรฒ richiedere molto tempo, quindi per eseguire la stessa operazione solo sui file modificati nel branch corrente, eseguire
```bash
make fixup
```
Quest'ultimo comando eseguirร anche tutti i controlli aggiuntivi per la consistenza del repository. Diamogli un'occhiata.
## Coerenza del repository
All'interno sono raggruppati tutti i test per assicurarsi che la tua PR lasci il repository in un buono stato ed รจ eseguito dal controllo `ci/circleci: check_repository_consistency`. Puoi eseguire localmente questo controllo eseguendo quanto segue:
```bash
make repo-consistency
```
Questo verifica che:
- Tutti gli oggetti aggiunti all'init sono documentati (eseguito da `utils/check_repo.py`)
- Tutti i file `__init__.py` hanno lo stesso contenuto nelle loro due sezioni (eseguito da `utils/check_inits.py`)
- Tutto il codice identificato come copia da un altro modulo รจ coerente con l'originale (eseguito da `utils/check_copies.py`)
- Le traduzioni dei README e l'indice della documentazione hanno lo stesso elenco di modelli del README principale (eseguito da `utils/check_copies.py`)
- Le tabelle autogenerate nella documentazione sono aggiornate (eseguito da `utils/check_table.py`)
- La libreria ha tutti gli oggetti disponibili anche se non tutte le dipendenze opzionali sono installate (eseguito da `utils/check_dummies.py`)
Se questo controllo fallisce, le prime due voci richiedono una correzione manuale, mentre le ultime quattro possono essere corrette automaticamente per te eseguendo il comando
```bash
make fix-copies
```
Ulteriori controlli riguardano le PR che aggiungono nuovi modelli, principalmente che:
- Tutti i modelli aggiunti sono in un Auto-mapping (eseguita da `utils/check_repo.py`)
<!-- TODO Sylvain, add a check that makes sure the common tests are implemented.-->
- Tutti i modelli sono testati correttamente (eseguito da `utils/check_repo.py`)
<!-- TODO Sylvain, add the following
- All models are added to the main README, inside the main doc
- All checkpoints used actually exist on the Hub
-->
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/add_new_pipeline.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Come creare una pipeline personalizzata?
In questa guida, scopriremo come creare una pipeline personalizzata e condividerla sull' [Hub](hf.co/models) o aggiungerla nella libreria
Transformers.
Innanzitutto, รจ necessario decidere gli input grezzi che la pipeline sarร in grado di accettare. Possono essere strings, raw bytes,
dictionaries o qualsiasi cosa sia l'input desiderato piรน probabile. Cerca di mantenere questi input il piรน possibile in Python
in quanto facilita la compatibilitร (anche con altri linguaggi tramite JSON). Questi saranno gli `inputs` della
pipeline (`preprocess`).
Poi definire gli `outputs`. Stessa strategia degli `inputs`. Piรน รจ seplice e meglio รจ. Questi saranno gli output del metodo
`postprocess`.
Si parte ereditando la classe base `Pipeline`. con i 4 metodi che bisogna implementare `preprocess`,
`_forward`, `postprocess` e `_sanitize_parameters`.
```python
from transformers import Pipeline
class MyPipeline(Pipeline):
def _sanitize_parameters(self, **kwargs):
preprocess_kwargs = {}
if "maybe_arg" in kwargs:
preprocess_kwargs["maybe_arg"] = kwargs["maybe_arg"]
return preprocess_kwargs, {}, {}
def preprocess(self, inputs, maybe_arg=2):
model_input = Tensor(inputs["input_ids"])
return {"model_input": model_input}
def _forward(self, model_inputs):
# model_inputs == {"model_input": model_input}
outputs = self.model(**model_inputs)
# Maybe {"logits": Tensor(...)}
return outputs
def postprocess(self, model_outputs):
best_class = model_outputs["logits"].softmax(-1)
return best_class
```
La struttura di questa suddivisione consiste nel supportare in modo relativamente continuo CPU/GPU, supportando allo stesso tempo l'esecuzione di
pre/postelaborazione sulla CPU su thread diversi.
`preprocess` prenderร gli input originariamente definiti e li trasformerร in qualcosa di alimentabile dal modello. Potrebbe
contenere piรน informazioni e di solito รจ un `Dict`.
`_forward` รจ il dettaglio dell'implementazione e non รจ destinato a essere chiamato direttamente. `forward` รจ il metodo preferito per assicurarsi che tutto funzioni correttamente perchรจ contiene delle slavaguardie. Se qualcosa รจ
รจ collegato a un modello reale, appartiene al metodo `_forward`, tutto il resto รจ nel preprocess/postprocess.
`postprocess` prende l'otput di `_forward` e lo trasforma nell'output finale che era stato deciso in precedenza.
`_sanitize_parameters` esiste per consentire agli utenti di passare i parametri ogni volta che desiderano sia a inizialization time `pipeline(...., maybe_arg=4)` che al call time `pipe = pipeline(...); output = pipe(...., maybe_arg=4)`.
`_sanitize_parameters` ritorna 3 dicts di kwargs che vengono passati direttamente a `preprocess`,
`_forward` e `postprocess`. Non riempire nulla se il chiamante non ha chiamato con alcun parametro aggiuntivo. Questo
consente di mantenere gli argomenti predefiniti nella definizione della funzione, che รจ sempre piรน "naturale".
Un esempio classico potrebbe essere l'argomento `top_k` nel post processing dei classification tasks.
```python
>>> pipe = pipeline("my-new-task")
>>> pipe("This is a test")
[{"label": "1-star", "score": 0.8}, {"label": "2-star", "score": 0.1}, {"label": "3-star", "score": 0.05}
{"label": "4-star", "score": 0.025}, {"label": "5-star", "score": 0.025}]
>>> pipe("This is a test", top_k=2)
[{"label": "1-star", "score": 0.8}, {"label": "2-star", "score": 0.1}]
```
In order to achieve that, we'll update our `postprocess` method with a default parameter to `5`. and edit
`_sanitize_parameters` to allow this new parameter.
```python
def postprocess(self, model_outputs, top_k=5):
best_class = model_outputs["logits"].softmax(-1)
# Add logic to handle top_k
return best_class
def _sanitize_parameters(self, **kwargs):
preprocess_kwargs = {}
if "maybe_arg" in kwargs:
preprocess_kwargs["maybe_arg"] = kwargs["maybe_arg"]
postprocess_kwargs = {}
if "top_k" in kwargs:
postprocess_kwargs["top_k"] = kwargs["top_k"]
return preprocess_kwargs, {}, postprocess_kwargs
```
Cercare di mantenere gli input/output molto semplici e idealmente serializzabili in JSON, in quanto ciรฒ rende l'uso della pipeline molto facile
senza richiedere agli utenti di comprendere nuovi tipi di oggetti. ร anche relativamente comune supportare molti tipi di argomenti
per facilitarne l'uso (ad esempio file audio, possono essere nomi di file, URL o byte puri).
## Aggiungilo alla lista dei tasks supportati
Per registrar il tuo `new-task` alla lista dei tasks supportati, devi aggiungerlo al `PIPELINE_REGISTRY`:
```python
from transformers.pipelines import PIPELINE_REGISTRY
PIPELINE_REGISTRY.register_pipeline(
"new-task",
pipeline_class=MyPipeline,
pt_model=AutoModelForSequenceClassification,
)
```
Puoi specificare il modello di default che desideri, in questo caso dovrebbe essere accompagnato da una revisione specifica (che puรฒ essere il nome di un branch o l'hash di un commit, in questo caso abbiamo preso `"abcdef"`) e anche dal type:
```python
PIPELINE_REGISTRY.register_pipeline(
"new-task",
pipeline_class=MyPipeline,
pt_model=AutoModelForSequenceClassification,
default={"pt": ("user/awesome_model", "abcdef")},
type="text", # current support type: text, audio, image, multimodal
)
```
## Condividi la tua pipeline sull'Hub
Per condividere la tua pipeline personalizzata sull'Hub, devi solo salvare il codice della tua sottoclasse `Pipeline` in un file
python. Per esempio, supponiamo di voler utilizzare una pipeline personalizzata per la classificazione delle coppie di frasi come la seguente:
```py
import numpy as np
from transformers import Pipeline
def softmax(outputs):
maxes = np.max(outputs, axis=-1, keepdims=True)
shifted_exp = np.exp(outputs - maxes)
return shifted_exp / shifted_exp.sum(axis=-1, keepdims=True)
class PairClassificationPipeline(Pipeline):
def _sanitize_parameters(self, **kwargs):
preprocess_kwargs = {}
if "second_text" in kwargs:
preprocess_kwargs["second_text"] = kwargs["second_text"]
return preprocess_kwargs, {}, {}
def preprocess(self, text, second_text=None):
return self.tokenizer(text, text_pair=second_text, return_tensors=self.framework)
def _forward(self, model_inputs):
return self.model(**model_inputs)
def postprocess(self, model_outputs):
logits = model_outputs.logits[0].numpy()
probabilities = softmax(logits)
best_class = np.argmax(probabilities)
label = self.model.config.id2label[best_class]
score = probabilities[best_class].item()
logits = logits.tolist()
return {"label": label, "score": score, "logits": logits}
```
L'implementazione รจ agnostica al framework, e lavorerร sia con modelli PyTorch che con TensorFlow. Se l'abbiamo salvato in un file chiamato `pair_classification.py`, puรฒ essere successivamente importato e registrato in questo modo:
```py
from pair_classification import PairClassificationPipeline
from transformers.pipelines import PIPELINE_REGISTRY
from transformers import AutoModelForSequenceClassification, TFAutoModelForSequenceClassification
PIPELINE_REGISTRY.register_pipeline(
"pair-classification",
pipeline_class=PairClassificationPipeline,
pt_model=AutoModelForSequenceClassification,
tf_model=TFAutoModelForSequenceClassification,
)
```
Una volta fatto, possiamo usarla con un modello pretrained. L'istanza `sgugger/finetuned-bert-mrpc` รจ stata
fine-tuned sul dataset MRPC, che classifica le coppie di frasi come parafrasi o no.
```py
from transformers import pipeline
classifier = pipeline("pair-classification", model="sgugger/finetuned-bert-mrpc")
```
Successivamente possiamo condividerlo sull'Hub usando il metodo `save_pretrained` in un `Repository`:
```py
from huggingface_hub import Repository
repo = Repository("test-dynamic-pipeline", clone_from="{your_username}/test-dynamic-pipeline")
classifier.save_pretrained("test-dynamic-pipeline")
repo.push_to_hub()
```
Questo codice copierร il file dove รจ stato definitp `PairClassificationPipeline` all'interno della cartella `"test-dynamic-pipeline"`,
insieme al salvataggio del modello e del tokenizer della pipeline, prima di pushare il tutto nel repository
`{your_username}/test-dynamic-pipeline`. Dopodichรฉ chiunque potrร utilizzarlo, purchรฉ fornisca l'opzione
`trust_remote_code=True`:
```py
from transformers import pipeline
classifier = pipeline(model="{your_username}/test-dynamic-pipeline", trust_remote_code=True)
```
## Aggiungere la pipeline a Transformers
Se vuoi contribuire con la tua pipeline a Transformers, dovrai aggiungere un modulo nel sottomodulo `pipelines`
con il codice della tua pipeline, quindi aggiungilo all'elenco dei tasks definiti in `pipelines/__init__.py`.
Poi hai bisogno di aggiungere i test. Crea un nuovo file `tests/test_pipelines_MY_PIPELINE.py` con esempi ed altri test.
La funzione `run_pipeline_test` sarร molto generica e su piccoli modelli casuali su ogni possibile
architettura, come definito da `model_mapping` e `tf_model_mapping`.
Questo รจ molto importante per testare la compatibilitร futura, nel senso che se qualcuno aggiunge un nuovo modello di
`XXXForQuestionAnswering` allora il test della pipeline tenterร di essere eseguito su di esso. Poichรฉ i modelli sono casuali, รจ
รจ impossibile controllare i valori effettivi, per questo esiste un aiuto `ANY` che tenterร solamente di far corrispondere l'output della pipeline TYPE.
Hai anche *bisogno* di implementare 2 (idealmente 4) test.
- `test_small_model_pt` : Definire 1 piccolo modello per questa pipeline (non importa se i risultati non hanno senso)
e testare i risultati della pipeline. I risultati dovrebbero essere gli stessi di `test_small_model_tf`.
- `test_small_model_tf` : Definire 1 piccolo modello per questa pipeline (non importa se i risultati non hanno senso)
e testare i risultati della pipeline. I risultati dovrebbero essere gli stessi di `test_small_model_pt`.
- `test_large_model_pt` (`optional`): Testare la pipeline su una pipeline reale in cui i risultati dovrebbero avere
senso. Questi test sono lenti e dovrebbero essere contrassegnati come tali. In questo caso l'obiettivo รจ mostrare la pipeline e assicurarsi che non ci siano derive nelle versioni future
- `test_large_model_tf` (`optional`): Testare la pipeline su una pipeline reale in cui i risultati dovrebbero avere
senso. Questi test sono lenti e dovrebbero essere contrassegnati come tali. In questo caso l'obiettivo รจ mostrare la pipeline e assicurarsi
che non ci siano derive nelle versioni future
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/perf_infer_gpu_many.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Inferenza Efficiente su GPU Multiple
Questo documento contiene informazioni su come fare inferenza in maniera efficiente su GPU multiple.
<Tip>
Nota: Un setup con GPU multiple puรฒ utilizzare la maggior parte delle strategie descritte nella [sezione con GPU singola](./perf_infer_gpu_one). Tuttavia, รจ necessario conoscere delle tecniche semplici che possono essere utilizzate per un risultato migliore.
</Tip>
## `BetterTransformer` per inferenza piรน rapida
Abbiamo recentemente integrato `BetterTransformer` per inferenza piรน rapida su multi-GPU per modelli su testo, immagini e audio. Controlla il documento con queste integrazioni [qui](https://huggingface.co/docs/optimum/bettertransformer/overview) per maggiori dettagli.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/community.md
|
<!--โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Comunitร
Questa pagina raggruppa le risorse sviluppate dalla comunitร riguardo ๐ค Transformers.
## Risorse della comunitร :
| Risorsa | Descrizione | Autore |
|:----------|:-------------|------:|
| [Glossario delle Flashcards di Transformers](https://www.darigovresearch.com/huggingface-transformers-glossary-flashcards) | Un insieme di flashcards basate sul [glossario della documentazione di Transformers](glossary), creato in un formato tale da permettere un facile apprendimento e revisione usando [Anki](https://apps.ankiweb.net/), un'applicazione open-source e multi-piattaforma, specificatamente progettata per ricordare informazioni nel lungo termine. Guarda questo [video introduttivo su come usare le flashcards](https://www.youtube.com/watch?v=Dji_h7PILrw). | [Darigov Research](https://www.darigovresearch.com/) |
## Notebook della comunitร :
| Notebook | Descrizione | Autore | |
|:----------|:-------------|:-------------|------:|
| [Fine-tuning di un Transformer pre-addestrato, al fine di generare testi di canzoni](https://github.com/AlekseyKorshuk/huggingartists) | Come generare testi di canzoni nello stile del vostro artista preferito attraverso il fine-tuning di un modello GPT-2. | [Aleksey Korshuk](https://github.com/AlekseyKorshuk) | [](https://colab.research.google.com/github/AlekseyKorshuk/huggingartists/blob/master/huggingartists-demo.ipynb) |
| [Addestramento di T5 in Tensorflow 2 ](https://github.com/snapthat/TF-T5-text-to-text) | Come addestrare T5 per qualsiasi attivitร usando Tensorflow 2. Questo notebook mostra come risolvere l'attivitร di "Question Answering" usando Tensorflow 2 e SQUAD. | [Muhammad Harris](https://github.com/HarrisDePerceptron) |[](https://colab.research.google.com/github/snapthat/TF-T5-text-to-text/blob/master/snapthatT5/notebooks/TF-T5-Datasets%20Training.ipynb) |
| [Addestramento di T5 con TPU](https://github.com/patil-suraj/exploring-T5/blob/master/T5_on_TPU.ipynb) | Come addestrare T5 su SQUAD con Transformers e NLP. | [Suraj Patil](https://github.com/patil-suraj) |[](https://colab.research.google.com/github/patil-suraj/exploring-T5/blob/master/T5_on_TPU.ipynb#scrollTo=QLGiFCDqvuil) |
| [Fine-tuning di T5 per la classificazione e scelta multipla](https://github.com/patil-suraj/exploring-T5/blob/master/t5_fine_tuning.ipynb) | Come effettuare il fine-tuning di T5 per le attivitร di classificazione a scelta multipla - usando un formato testo-a-testo - con PyTorch Lightning. | [Suraj Patil](https://github.com/patil-suraj) | [](https://colab.research.google.com/github/patil-suraj/exploring-T5/blob/master/t5_fine_tuning.ipynb) |
| [Fine-tuning di DialoGPT su nuovi dataset e lingue](https://github.com/ncoop57/i-am-a-nerd/blob/master/_notebooks/2020-05-12-chatbot-part-1.ipynb) | Come effettuare il fine-tuning di un modello DialoGPT su un nuovo dataset per chatbots conversazionali open-dialog. | [Nathan Cooper](https://github.com/ncoop57) | [](https://colab.research.google.com/github/ncoop57/i-am-a-nerd/blob/master/_notebooks/2020-05-12-chatbot-part-1.ipynb) |
| [Modellamento di una lunga sequenza con Reformer](https://github.com/patrickvonplaten/notebooks/blob/master/PyTorch_Reformer.ipynb) | Come addestrare su sequenze di lunghezza fino a 500 mila token con Reformer. | [Patrick von Platen](https://github.com/patrickvonplaten) | [](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/PyTorch_Reformer.ipynb) |
| [Fine-tuning di BART per riassumere testi](https://github.com/ohmeow/ohmeow_website/blob/master/_notebooks/2020-05-23-text-generation-with-blurr.ipynb) | Come effettuare il fine-tuning di BART per riassumere testi con fastai usando blurr. | [Wayde Gilliam](https://ohmeow.com/) | [](https://colab.research.google.com/github/ohmeow/ohmeow_website/blob/master/_notebooks/2020-05-23-text-generation-with-blurr.ipynb) |
| [Fine-tuning di un Transformer pre-addestrato su tweet](https://colab.research.google.com/github/borisdayma/huggingtweets/blob/master/huggingtweets-demo.ipynb) | Come generare tweet nello stile del tuo account Twitter preferito attraverso il fine-tuning di un modello GPT-2. | [Boris Dayma](https://github.com/borisdayma) | [](https://colab.research.google.com/github/borisdayma/huggingtweets/blob/master/huggingtweets-demo.ipynb) |
| [Ottimizzazione di modelli ๐ค Hugging Face con Weights & Biases](https://colab.research.google.com/github/wandb/examples/blob/master/colabs/huggingface/Optimize_Hugging_Face_models_with_Weights_%26_Biases.ipynb) | Un tutorial completo che mostra l'integrazione di W&B con Hugging Face. | [Boris Dayma](https://github.com/borisdayma) | [](https://colab.research.google.com/github/wandb/examples/blob/master/colabs/huggingface/Optimize_Hugging_Face_models_with_Weights_%26_Biases.ipynb) |
| [Longformer pre-addestrato](https://github.com/allenai/longformer/blob/master/scripts/convert_model_to_long.ipynb) | Come costruire una versione "long" degli esistenti modelli pre-addestrati. | [Iz Beltagy](https://beltagy.net) | [](https://colab.research.google.com/github/allenai/longformer/blob/master/scripts/convert_model_to_long.ipynb) |
| [Fine-tuning di Longformer per QA](https://github.com/patil-suraj/Notebooks/blob/master/longformer_qa_training.ipynb) | Come effettuare il fine-tuning di un modello longformer per un task di QA.| [Suraj Patil](https://github.com/patil-suraj) | [](https://colab.research.google.com/github/patil-suraj/Notebooks/blob/master/longformer_qa_training.ipynb) |
| [Valutazione di modelli con ๐คNLP](https://github.com/patrickvonplaten/notebooks/blob/master/How_to_evaluate_Longformer_on_TriviaQA_using_NLP.ipynb) | Come valutare longformer su TriviaQA con `NLP`. | [Patrick von Platen](https://github.com/patrickvonplaten) | [](https://colab.research.google.com/drive/1m7eTGlPmLRgoPkkA7rkhQdZ9ydpmsdLE?usp=sharing) |
| [Fine-tuning di T5 per Sentiment Span Extraction](https://github.com/enzoampil/t5-intro/blob/master/t5_qa_training_pytorch_span_extraction.ipynb) | Come effettuare il fine-tuning di T5 per la sentiment span extraction - usando un formato testo-a-testo - con PyTorch Lightning. | [Lorenzo Ampil](https://github.com/enzoampil) | [](https://colab.research.google.com/github/enzoampil/t5-intro/blob/master/t5_qa_training_pytorch_span_extraction.ipynb) |
| [Fine-tuning di DistilBert per la classificazione multi-classe](https://github.com/abhimishra91/transformers-tutorials/blob/master/transformers_multiclass_classification.ipynb) | Come effettuare il fine-tuning di DistilBert per la classificazione multi-classe con PyTorch. | [Abhishek Kumar Mishra](https://github.com/abhimishra91) | [](https://colab.research.google.com/github/abhimishra91/transformers-tutorials/blob/master/transformers_multiclass_classification.ipynb)|
|[Fine-tuning di BERT per la classificazione multi-etichetta](https://github.com/abhimishra91/transformers-tutorials/blob/master/transformers_multi_label_classification.ipynb)|Come effettuare il fine-tuning di BERT per la classificazione multi-etichetta con PyTorch. |[Abhishek Kumar Mishra](https://github.com/abhimishra91) |[](https://colab.research.google.com/github/abhimishra91/transformers-tutorials/blob/master/transformers_multi_label_classification.ipynb)|
|[Accelerazione del fine-tuning con il Dynamic Padding / Bucketing](https://github.com/ELS-RD/transformers-notebook/blob/master/Divide_Hugging_Face_Transformers_training_time_by_2_or_more.ipynb)| Come velocizzare il fine-tuning di un fattore 2X usando il dynamic padding / bucketing. |[Michael Benesty](https://github.com/pommedeterresautee) |[](https://colab.research.google.com/drive/1CBfRU1zbfu7-ijiOqAAQUA-RJaxfcJoO?usp=sharing)|
|[Pre-addestramento di Reformer per Masked Language Modeling](https://github.com/patrickvonplaten/notebooks/blob/master/Reformer_For_Masked_LM.ipynb)| Come addestrare un modello Reformer usando livelli di self-attention bi-direzionali.| [Patrick von Platen](https://github.com/patrickvonplaten) | [](https://colab.research.google.com/drive/1tzzh0i8PgDQGV3SMFUGxM7_gGae3K-uW?usp=sharing)|
|[Espansione e fine-tuning di Sci-BERT](https://github.com/lordtt13/word-embeddings/blob/master/COVID-19%20Research%20Data/COVID-SciBERT.ipynb)| Come incrementare il vocabolario di un modello SciBERT - pre-addestrato da AllenAI sul dataset CORD - e crearne una pipeline. | [Tanmay Thakur](https://github.com/lordtt13) | [](https://colab.research.google.com/drive/1rqAR40goxbAfez1xvF3hBJphSCsvXmh8)|
|[Fine-tuning di BlenderBotSmall per riassumere testi usando Trainer API](https://github.com/lordtt13/transformers-experiments/blob/master/Custom%20Tasks/fine-tune-blenderbot_small-for-summarization.ipynb)| Come effettuare il fine-tuning di BlenderBotSmall per riassumere testi su un dataset personalizzato, usando Trainer API. | [Tanmay Thakur](https://github.com/lordtt13) | [](https://colab.research.google.com/drive/19Wmupuls7mykSGyRN_Qo6lPQhgp56ymq?usp=sharing)|
|[Fine-tuning di Electra e interpretazione con Integrated Gradients](https://github.com/elsanns/xai-nlp-notebooks/blob/master/electra_fine_tune_interpret_captum_ig.ipynb) | Come effettuare il fine-tuning di Electra per l'analisi dei sentimenti e intepretare le predizioni con Captum Integrated Gradients. | [Eliza Szczechla](https://elsanns.github.io) | [](https://colab.research.google.com/github/elsanns/xai-nlp-notebooks/blob/master/electra_fine_tune_interpret_captum_ig.ipynb)|
|[Fine-tuning di un modello GPT-2 non inglese con la classe Trainer](https://github.com/philschmid/fine-tune-GPT-2/blob/master/Fine_tune_a_non_English_GPT_2_Model_with_Huggingface.ipynb) | Come effettuare il fine-tuning di un modello GPT-2 non inglese con la classe Trainer. | [Philipp Schmid](https://www.philschmid.de) | [](https://colab.research.google.com/github/philschmid/fine-tune-GPT-2/blob/master/Fine_tune_a_non_English_GPT_2_Model_with_Huggingface.ipynb)|
|[Fine-tuning di un modello DistilBERT per la classficazione multi-etichetta](https://github.com/DhavalTaunk08/Transformers_scripts/blob/master/Transformers_multilabel_distilbert.ipynb) | Come effettuare il fine-tuning di un modello DistilBERT per l'attivitร di classificazione multi-etichetta. | [Dhaval Taunk](https://github.com/DhavalTaunk08) | [](https://colab.research.google.com/github/DhavalTaunk08/Transformers_scripts/blob/master/Transformers_multilabel_distilbert.ipynb)|
|[Fine-tuning di ALBERT per la classifcazione di coppie di frasi](https://github.com/NadirEM/nlp-notebooks/blob/master/Fine_tune_ALBERT_sentence_pair_classification.ipynb) | Come effettuare il fine-tuning di un modello ALBERT - o un altro modello BERT-based - per l'attivitร di classificazione di coppie di frasi. | [Nadir El Manouzi](https://github.com/NadirEM) | [](https://colab.research.google.com/github/NadirEM/nlp-notebooks/blob/master/Fine_tune_ALBERT_sentence_pair_classification.ipynb)|
|[Fine-tuning di Roberta per l'analisi di sentimenti](https://github.com/DhavalTaunk08/NLP_scripts/blob/master/sentiment_analysis_using_roberta.ipynb) | Come effettuare il fine-tuning di un modello Roberta per l'analisi di sentimenti. | [Dhaval Taunk](https://github.com/DhavalTaunk08) | [](https://colab.research.google.com/github/DhavalTaunk08/NLP_scripts/blob/master/sentiment_analysis_using_roberta.ipynb)|
|[Valutazione di modelli che generano domande](https://github.com/flexudy-pipe/qugeev) | Quanto sono accurante le risposte alle domande generate dal tuo modello transformer seq2seq? | [Pascal Zoleko](https://github.com/zolekode) | [](https://colab.research.google.com/drive/1bpsSqCQU-iw_5nNoRm_crPq6FRuJthq_?usp=sharing)|
|[Classificazione di testo con DistilBERT e Tensorflow](https://github.com/peterbayerle/huggingface_notebook/blob/main/distilbert_tf.ipynb) | Come effettuare il fine-tuning di DistilBERT per la classificazione di testo in TensorFlow. | [Peter Bayerle](https://github.com/peterbayerle) | [](https://colab.research.google.com/github/peterbayerle/huggingface_notebook/blob/main/distilbert_tf.ipynb)|
|[Utilizzo di BERT per riassumere testi con un modello Encoder-Decoder su CNN/Dailymail](https://github.com/patrickvonplaten/notebooks/blob/master/BERT2BERT_for_CNN_Dailymail.ipynb) | Come avviare "a caldo" un *EncoderDecoderModel* attraverso l'utilizzo di un checkpoint *bert-base-uncased* per riassumere testi su CNN/Dailymail. | [Patrick von Platen](https://github.com/patrickvonplaten) | [](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/BERT2BERT_for_CNN_Dailymail.ipynb)|
|[Utilizzo di RoBERTa per riassumere testi con un modello Encoder-Decoder su BBC XSum](https://github.com/patrickvonplaten/notebooks/blob/master/RoBERTaShared_for_BBC_XSum.ipynb) | Come avviare "a caldo" un *EncoderDecoderModel* (condiviso) attraverso l'utilizzo di un checkpoint *roberta-base* per riassumere testi su BBC/XSum. | [Patrick von Platen](https://github.com/patrickvonplaten) | [](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/RoBERTaShared_for_BBC_XSum.ipynb)|
|[Fine-tuning di TAPAS su Sequential Question Answering (SQA)](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/TAPAS/Fine_tuning_TapasForQuestionAnswering_on_SQA.ipynb) | Come effettuare il fine-tuning di un modello *TapasForQuestionAnswering* attraverso l'utilizzo di un checkpoint *tapas-base* sul dataset Sequential Question Answering (SQA). | [Niels Rogge](https://github.com/nielsrogge) | [](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/TAPAS/Fine_tuning_TapasForQuestionAnswering_on_SQA.ipynb)|
|[Valutazione di TAPAS su Table Fact Checking (TabFact)](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/TAPAS/Evaluating_TAPAS_on_the_Tabfact_test_set.ipynb) | Come valutare un modello *TapasForSequenceClassification* - fine-tuned con un checkpoint *tapas-base-finetuned-tabfact* - usando una combinazione delle librerie ๐ค datasets e ๐ค transformers. | [Niels Rogge](https://github.com/nielsrogge) | [](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/TAPAS/Evaluating_TAPAS_on_the_Tabfact_test_set.ipynb)|
|[Fine-tuning di mBART per la traduzione](https://colab.research.google.com/github/vasudevgupta7/huggingface-tutorials/blob/main/translation_training.ipynb) | Come effettuare il fine-tuning di mBART usando Seq2SeqTrainer per la traduzione da hindi a inglese.| [Vasudev Gupta](https://github.com/vasudevgupta7) | [](https://colab.research.google.com/github/vasudevgupta7/huggingface-tutorials/blob/main/translation_training.ipynb)|
|[Fine-tuning di LayoutLM su FUNSD (un dataset per la comprensione della forma)](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/LayoutLM/Fine_tuning_LayoutLMForTokenClassification_on_FUNSD.ipynb) | Come effettuare il fine-tuning di un modello *LayoutLMForTokenClassification* sul dataset FUNSD per l'estrazione di informazioni da documenti scannerizzati.| [Niels Rogge](https://github.com/nielsrogge) | [](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/LayoutLM/Fine_tuning_LayoutLMForTokenClassification_on_FUNSD.ipynb)|
|[Fine-tuning di DistilGPT2 e generazione di testo](https://colab.research.google.com/github/tripathiaakash/DistilGPT2-Tutorial/blob/main/distilgpt2_fine_tuning.ipynb) | Come effettuare il fine-tuning di DistilGPT2 e generare testo. | [Aakash Tripathi](https://github.com/tripathiaakash) | [](https://colab.research.google.com/github/tripathiaakash/DistilGPT2-Tutorial/blob/main/distilgpt2_fine_tuning.ipynb)|
|[Fine-tuning di LED fino a 8 mila token](https://github.com/patrickvonplaten/notebooks/blob/master/Fine_tune_Longformer_Encoder_Decoder_(LED)_for_Summarization_on_pubmed.ipynb) | Come effettuare il fine-tuning di LED su PubMed per riassumere "lunghi" testi. | [Patrick von Platen](https://github.com/patrickvonplaten) | [](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/Fine_tune_Longformer_Encoder_Decoder_(LED)_for_Summarization_on_pubmed.ipynb)|
|[Valutazione di LED su Arxiv](https://github.com/patrickvonplaten/notebooks/blob/master/LED_on_Arxiv.ipynb) | Come valutare efficacemente LED sull'attivitร di riassumere "lunghi" testi. | [Patrick von Platen](https://github.com/patrickvonplaten) | [](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/LED_on_Arxiv.ipynb)|
|[Fine-tuning di LayoutLM su RVL-CDIP, un dataset per la classificazione di documenti (immagini)](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/LayoutLM/Fine_tuning_LayoutLMForSequenceClassification_on_RVL_CDIP.ipynb) | Come effettuare il fine-tuning di un modello *LayoutLMForSequenceClassification* sul dataset RVL-CDIP per la classificazione di documenti scannerizzati. | [Niels Rogge](https://github.com/nielsrogge) | [](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/LayoutLM/Fine_tuning_LayoutLMForSequenceClassification_on_RVL_CDIP.ipynb)|
|[Decodifica Wav2Vec2 CTC con variazioni di GPT2](https://github.com/voidful/huggingface_notebook/blob/main/xlsr_gpt.ipynb) | Come decodificare sequenze CTC, variate da modelli di linguaggio. | [Eric Lam](https://github.com/voidful) | [](https://colab.research.google.com/drive/1e_z5jQHYbO2YKEaUgzb1ww1WwiAyydAj?usp=sharing)
|[Fine-tuning di BART per riassumere testi in due lingue con la classe Trainer](https://github.com/elsanns/xai-nlp-notebooks/blob/master/fine_tune_bart_summarization_two_langs.ipynb) | Come effettuare il fine-tuning di BART per riassumere testi in due lingue usando la classe Trainer. | [Eliza Szczechla](https://github.com/elsanns) | [](https://colab.research.google.com/github/elsanns/xai-nlp-notebooks/blob/master/fine_tune_bart_summarization_two_langs.ipynb)|
|[Valutazione di Big Bird su Trivia QA](https://github.com/patrickvonplaten/notebooks/blob/master/Evaluating_Big_Bird_on_TriviaQA.ipynb) | Come valutare BigBird su question answering di "lunghi" documenti attraverso Trivia QA. | [Patrick von Platen](https://github.com/patrickvonplaten) | [](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/Evaluating_Big_Bird_on_TriviaQA.ipynb)|
| [Creazione di sottotitoli per video usando Wav2Vec2](https://github.com/Muennighoff/ytclipcc/blob/main/wav2vec_youtube_captions.ipynb) | Come creare sottotitoli per qualsiasi video di YouTube trascrivendo l'audio con Wav2Vec. | [Niklas Muennighoff](https://github.com/Muennighoff) |[](https://colab.research.google.com/github/Muennighoff/ytclipcc/blob/main/wav2vec_youtube_captions.ipynb) |
| [Fine-tuning di Vision Transformer su CIFAR-10 usando PyTorch Lightning](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_PyTorch_Lightning.ipynb) | Come effettuare il fine-tuning di Vision Transformer (ViT) su CIFAR-10 usando HuggingFace Transformers, Datasets e PyTorch Lightning.| [Niels Rogge](https://github.com/nielsrogge) |[](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_PyTorch_Lightning.ipynb) |
| [Fine-tuning di Vision Transformer su CIFAR-10 usando ๐ค Trainer](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_the_%F0%9F%A4%97_Trainer.ipynb) | Come effettuare il fine-tuning di Vision Transformer (ViT) su CIFAR-10 usando HuggingFace Transformers, Datasets e ๐ค Trainer. | [Niels Rogge](https://github.com/nielsrogge) |[](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/VisionTransformer/Fine_tuning_the_Vision_Transformer_on_CIFAR_10_with_the_%F0%9F%A4%97_Trainer.ipynb) |
| [Valutazione di LUKE su Open Entity, un dataset di entity typing](https://github.com/studio-ousia/luke/blob/master/notebooks/huggingface_open_entity.ipynb) | Come valutare un modello *LukeForEntityClassification* sul dataset Open Entity. | [Ikuya Yamada](https://github.com/ikuyamada) |[](https://colab.research.google.com/github/studio-ousia/luke/blob/master/notebooks/huggingface_open_entity.ipynb) |
| [Valutazione di LUKE su TACRED, un dataset per l'estrazione di relazioni](https://github.com/studio-ousia/luke/blob/master/notebooks/huggingface_tacred.ipynb) | Come valutare un modello *LukeForEntityPairClassification* sul dataset TACRED. | [Ikuya Yamada](https://github.com/ikuyamada) |[](https://colab.research.google.com/github/studio-ousia/luke/blob/master/notebooks/huggingface_tacred.ipynb) |
| [Valutazione di LUKE su CoNLL-2003, un importante benchmark NER](https://github.com/studio-ousia/luke/blob/master/notebooks/huggingface_conll_2003.ipynb) | Come valutare un modello *LukeForEntitySpanClassification* sul dataset CoNLL-2003. | [Ikuya Yamada](https://github.com/ikuyamada) |[](https://colab.research.google.com/github/studio-ousia/luke/blob/master/notebooks/huggingface_conll_2003.ipynb) |
| [Valutazione di BigBird-Pegasus su dataset PubMed](https://github.com/vasudevgupta7/bigbird/blob/main/notebooks/bigbird_pegasus_evaluation.ipynb) | Come valutare un modello *BigBirdPegasusForConditionalGeneration* su dataset PubMed. | [Vasudev Gupta](https://github.com/vasudevgupta7) | [](https://colab.research.google.com/github/vasudevgupta7/bigbird/blob/main/notebooks/bigbird_pegasus_evaluation.ipynb) |
| [Classificazione di emozioni dal discorso con Wav2Vec2](https://github/m3hrdadfi/soxan/blob/main/notebooks/Emotion_recognition_in_Greek_speech_using_Wav2Vec2.ipynb) | Come utilizzare un modello pre-addestrato Wav2Vec2 per la classificazione di emozioni sul dataset MEGA. | [Mehrdad Farahani](https://github.com/m3hrdadfi) | [](https://colab.research.google.com/github/m3hrdadfi/soxan/blob/main/notebooks/Emotion_recognition_in_Greek_speech_using_Wav2Vec2.ipynb) |
| [Rilevamento oggetti in un'immagine con DETR](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/DETR/DETR_minimal_example_(with_DetrFeatureExtractor).ipynb) | Come usare un modello addestrato *DetrForObjectDetection* per rilevare oggetti in un'immagine e visualizzare l'attention. | [Niels Rogge](https://github.com/NielsRogge) | [](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/DETR/DETR_minimal_example_(with_DetrFeatureExtractor).ipynb) |
| [Fine-tuning di DETR su un dataset personalizzato per rilevare oggetti](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/DETR/Fine_tuning_DetrForObjectDetection_on_custom_dataset_(balloon).ipynb) | Come effettuare fine-tuning di un modello *DetrForObjectDetection* su un dataset personalizzato per rilevare oggetti. | [Niels Rogge](https://github.com/NielsRogge) | [](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/DETR/Fine_tuning_DetrForObjectDetection_on_custom_dataset_(balloon).ipynb) |
| [Fine-tuning di T5 per Named Entity Recognition](https://github.com/ToluClassics/Notebooks/blob/main/T5_Ner_Finetuning.ipynb) | Come effettuare fine-tunining di *T5* per un'attivitร di Named Entity Recognition. | [Ogundepo Odunayo](https://github.com/ToluClassics) | [](https://colab.research.google.com/drive/1obr78FY_cBmWY5ODViCmzdY6O1KB65Vc?usp=sharing) |
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/perf_infer_gpu_one.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Inferenza efficiente su GPU singola
Questo documento sarร presto completato con informazioni su come effetture l'inferenza su una singola GPU. Nel frattempo รจ possibile consultare [la guida per l'addestramento su una singola GPU](perf_train_gpu_one) e [la guida per l'inferenza su CPU](perf_infer_cpu).
## `BetterTransformer` per l'inferenza piรน veloce
Abbiamo recentemente integrato `BetterTransformer` per velocizzare l'inferenza su GPU per modelli di testo, immagini e audio. Per maggiori dettagli, consultare la documentazione su questa integrazione [qui](https://huggingface.co/docs/optimum/bettertransformer/overview).
## Integrazione di `bitsandbytes` per Int8 mixed-precision matrix decomposition
<Tip>
Nota che questa funzione puรฒ essere utilizzata anche nelle configurazioni multi GPU.
</Tip>
Dal paper [`LLM.int8() : 8-bit Matrix Multiplication for Transformers at Scale`](https://arxiv.org/abs/2208.07339), noi supportiamo l'integrazione di Hugging Face per tutti i modelli dell'Hub con poche righe di codice.
Il metodo `nn.Linear` riduce la dimensione di 2 per i pesi `float16` e `bfloat16` e di 4 per i pesi `float32`, con un impatto quasi nullo sulla qualitร , operando sugli outlier in half-precision.
Il metodo Int8 mixed-precision matrix decomposition funziona separando la moltiplicazione tra matrici in due flussi: (1) una matrice di flusso di outlier di caratteristiche sistematiche moltiplicata in fp16, (2) in flusso regolare di moltiplicazione di matrici int8 (99,9%). Con questo metodo, รจ possibile effettutare inferenza int8 per modelli molto grandi senza degrado predittivo.
Per maggiori dettagli sul metodo, consultare il [paper](https://arxiv.org/abs/2208.07339) o il nostro [blogpost sull'integrazione](https://huggingface.co/blog/hf-bitsandbytes-integration).
Nota che รจ necessaria una GPU per eseguire modelli di tipo mixed-8bit, poichรฉ i kernel sono stati compilati solo per le GPU. Prima di utilizzare questa funzione, assicurarsi di disporre di memoria sufficiente sulla GPU per memorizzare un quarto del modello (o la metร se i pesi del modello sono in mezza precisione).
Di seguito sono riportate alcune note per aiutarvi a utilizzare questo modulo, oppure seguite le dimostrazioni su [Google colab](#colab-demos).
### Requisiti
- Se si dispone di `bitsandbytes<0.37.0`, assicurarsi di eseguire su GPU NVIDIA che supportano tensor cores a 8 bit (Turing, Ampere o architetture piรน recenti - ad esempio T4, RTX20s RTX30s, A40-A100). Per `bitsandbytes>=0.37.0`, tutte le GPU dovrebbero essere supportate.
- Installare la versione corretta di `bitsandbytes` eseguendo:
`pip install bitsandbytes>=0.31.5`.
- Installare `accelerate`
`pip install accelerate>=0.12.0`
### Esecuzione di modelli mixed-Int8 - configurazione per singola GPU
Dopo aver installato le librerie necessarie, per caricare il tuo modello mixed 8-bit รจ il seguente:
```py
from transformers import AutoModelForCausalLM
model_name = "bigscience/bloom-2b5"
model_8bit = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", load_in_8bit=True)
```
Per la generazione di testo, si consiglia di:
* utilizzare il metodo `generate()` del modello invece della funzione `pipeline()`. Sebbene l'inferenza sia possibile con la funzione `pipeline()`, essa non รจ ottimizzata per i modelli mixed-8bit e sarร piรน lenta rispetto all'uso del metodo `generate()`. Inoltre, alcune strategie di campionamento, come il campionamento nucleaus, non sono supportate dalla funzione `pipeline()` per i modelli mixed-8bit.
* collocare tutti gli ingressi sullo stesso dispositivo del modello.
Ecco un semplice esempio:
```py
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "bigscience/bloom-2b5"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model_8bit = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", load_in_8bit=True)
text = "Hello, my llama is cute"
inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
generated_ids = model.generate(**inputs)
outputs = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
```
### Esecuzione di modelli mixed-8bit - configurazione multi GPU
Usare il seguente modo caricare il modello mixed-8bit su piรน GPU (stesso comando della configurazione a GPU singola):
```py
model_name = "bigscience/bloom-2b5"
model_8bit = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto", load_in_8bit=True)
```
Puoi controllare la RAM della GPU che si vuole allocare su ogni GPU usando `accelerate`. Utilizzare l'argomento `max_memory` come segue:
```py
max_memory_mapping = {0: "1GB", 1: "2GB"}
model_name = "bigscience/bloom-3b"
model_8bit = AutoModelForCausalLM.from_pretrained(
model_name, device_map="auto", load_in_8bit=True, max_memory=max_memory_mapping
)
```
In questo esempio, la prima GPU utilizzerร 1 GB di memoria e la seconda 2 GB.
### Colab demos
Con questo metodo รจ possibile inferire modelli che prima non era possibile inferire su Google Colab.
Guardate la demo per l'esecuzione di T5-11b (42GB in fp32)! Utilizzo la quantizzazione a 8 bit su Google Colab:
[](https://colab.research.google.com/drive/1YORPWx4okIHXnjW7MSAidXN29mPVNT7F?usp=sharing)
Oppure questa demo di BLOOM-3B:
[](https://colab.research.google.com/drive/1qOjXfQIAULfKvZqwCen8-MoWKGdSatZ4?usp=sharing)
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/training.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Fine-tuning di un modello pre-addestrato
[[open-in-colab]]
Ci sono benefici significativi nell'usare un modello pre-addestrato. Si riducono i costi computazionali, l'impronta di carbonio e ti consente di usare modelli stato dell'arte senza doverli addestrare da zero. ๐ค Transformers consente l'accesso a migliaia di modelli pre-addestrati per un'ampia gamma di compiti. Quando usi un modello pre-addestrato, lo alleni su un dataset specifico per il tuo compito. Questo รจ conosciuto come fine-tuning, una tecnica di addestramento incredibilmente potente. In questa esercitazione, potrai fare il fine-tuning di un modello pre-addestrato, con un framework di deep learning a tua scelta:
* Fine-tuning di un modello pre-addestrato con ๐ค Transformers [`Trainer`].
* Fine-tuning di un modello pre-addestrato in TensorFlow con Keras.
* Fine-tuning di un modello pre-addestrato con PyTorch.
<a id='data-processing'></a>
## Preparare un dataset
<Youtube id="_BZearw7f0w"/>
Prima di poter fare il fine-tuning di un modello pre-addestrato, scarica un dataset e preparalo per l'addestramento. La precedente esercitazione ti ha mostrato come processare i dati per l'addestramento e adesso hai l'opportunitร di metterti alla prova!
Inizia caricando il dataset [Yelp Reviews](https://huggingface.co/datasets/yelp_review_full):
```py
>>> from datasets import load_dataset
>>> dataset = load_dataset("yelp_review_full")
>>> dataset["train"][100]
{'label': 0,
'text': 'My expectations for McDonalds are t rarely high. But for one to still fail so spectacularly...that takes something special!\\nThe cashier took my friends\'s order, then promptly ignored me. I had to force myself in front of a cashier who opened his register to wait on the person BEHIND me. I waited over five minutes for a gigantic order that included precisely one kid\'s meal. After watching two people who ordered after me be handed their food, I asked where mine was. The manager started yelling at the cashiers for \\"serving off their orders\\" when they didn\'t have their food. But neither cashier was anywhere near those controls, and the manager was the one serving food to customers and clearing the boards.\\nThe manager was rude when giving me my order. She didn\'t make sure that I had everything ON MY RECEIPT, and never even had the decency to apologize that I felt I was getting poor service.\\nI\'ve eaten at various McDonalds restaurants for over 30 years. I\'ve worked at more than one location. I expect bad days, bad moods, and the occasional mistake. But I have yet to have a decent experience at this store. It will remain a place I avoid unless someone in my party needs to avoid illness from low blood sugar. Perhaps I should go back to the racially biased service of Steak n Shake instead!'}
```
Come giร sai, hai bisogno di un tokenizer per processare il testo e includere una strategia di padding e truncation per gestire sequenze di lunghezza variabile. Per processare il dataset in un unico passo, usa il metodo [`map`](https://huggingface.co/docs/datasets/process#map) di ๐ค Datasets che applica la funzione di preprocessing all'intero dataset:
```py
>>> from transformers import AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("bert-base-cased")
>>> def tokenize_function(examples):
... return tokenizer(examples["text"], padding="max_length", truncation=True)
>>> tokenized_datasets = dataset.map(tokenize_function, batched=True)
```
Se vuoi, puoi creare un sottoinsieme piรน piccolo del dataset per il fine-tuning cosรฌ da ridurre il tempo necessario:
```py
>>> small_train_dataset = tokenized_datasets["train"].shuffle(seed=42).select(range(1000))
>>> small_eval_dataset = tokenized_datasets["test"].shuffle(seed=42).select(range(1000))
```
<a id='trainer'></a>
## Addestramento
<frameworkcontent>
<pt>
<Youtube id="nvBXf7s7vTI"/>
๐ค Transformers mette a disposizione la classe [`Trainer`] ottimizzata per addestrare modelli ๐ค Transformers, rendendo semplice iniziare l'addestramento senza scrivere manualmente il tuo ciclo di addestramento. L'API [`Trainer`] supporta un'ampia gamma di opzioni e funzionalitร di addestramento come logging, gradient accumulation e mixed precision.
Inizia caricando il tuo modello e specificando il numero di etichette (labels) attese. Nel dataset Yelp Review [dataset card](https://huggingface.co/datasets/yelp_review_full#data-fields), sai che ci sono cinque etichette:
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained("bert-base-cased", num_labels=5)
```
<Tip>
Potresti vedere un warning dato che alcuni dei pesi pre-addestrati non sono stati utilizzati e altri pesi sono stati inizializzati casualmente. Non preoccuparti, รจ completamente normale! L'head pre-addestrata del modello BERT viene scartata e rimpiazzata da una classification head inizializzata casualmente. Farai il fine-tuning di questa nuova head del modello sul tuo compito di classificazione, trasferendogli la conoscenza del modello pre-addestrato.
</Tip>
### Iperparametri per il training
Successivamente, crea una classe [`TrainingArguments`] contenente tutti gli iperparametri che si possono regore nonchรฉ le variabili per attivare le differenti opzioni di addestramento. Per questa esercitazione puoi iniziare con gli [iperparametri](https://huggingface.co/docs/transformers/main_classes/trainer#transformers.TrainingArguments) di ddestramento predefiniti, ma sentiti libero di sperimentare per trovare la configurazione ottimale per te.
Specifica dove salvare i checkpoints del tuo addestramento:
```py
>>> from transformers import TrainingArguments
>>> training_args = TrainingArguments(output_dir="test_trainer")
```
### Metriche
[`Trainer`] non valuta automaticamente le performance del modello durante l'addestramento. Dovrai passare a [`Trainer`] una funzione che calcola e restituisce le metriche. La libreria ๐ค Datasets mette a disposizione una semplice funzione [`accuracy`](https://huggingface.co/metrics/accuracy) che puoi caricare con la funzione `load_metric` (guarda questa [esercitazione](https://huggingface.co/docs/datasets/metrics) per maggiori informazioni):
```py
>>> import numpy as np
>>> from datasets import load_metric
>>> metric = load_metric("accuracy")
```
Richiama `compute` su `metric` per calcolare l'accuratezza delle tue previsioni. Prima di passare le tue previsioni a `compute`, hai bisogno di convertirle in logits (ricorda che tutti i modelli ๐ค Transformers restituiscono logits):
```py
>>> def compute_metrics(eval_pred):
... logits, labels = eval_pred
... predictions = np.argmax(logits, axis=-1)
... return metric.compute(predictions=predictions, references=labels)
```
Se preferisci monitorare le tue metriche di valutazione durante il fine-tuning, specifica il parametro `evaluation_strategy` nei tuoi training arguments per restituire le metriche di valutazione ad ogni epoca di addestramento:
```py
>>> from transformers import TrainingArguments, Trainer
>>> training_args = TrainingArguments(output_dir="test_trainer", evaluation_strategy="epoch")
```
### Trainer
Crea un oggetto [`Trainer`] col tuo modello, training arguments, dataset di training e test, e funzione di valutazione:
```py
>>> trainer = Trainer(
... model=model,
... args=training_args,
... train_dataset=small_train_dataset,
... eval_dataset=small_eval_dataset,
... compute_metrics=compute_metrics,
... )
```
Poi metti a punto il modello richiamando [`~transformers.Trainer.train`]:
```py
>>> trainer.train()
```
</pt>
<tf>
<a id='keras'></a>
<Youtube id="rnTGBy2ax1c"/>
I modelli ๐ค Transformers supportano anche l'addestramento in TensorFlow usando l'API di Keras.
### Convertire dataset nel formato per TensorFlow
Il [`DefaultDataCollator`] assembla tensori in lotti su cui il modello si addestrerร . Assicurati di specificare di restituire tensori per TensorFlow in `return_tensors`:
```py
>>> from transformers import DefaultDataCollator
>>> data_collator = DefaultDataCollator(return_tensors="tf")
```
<Tip>
[`Trainer`] usa [`DataCollatorWithPadding`] in maniera predefinita in modo da non dover specificare esplicitamente un collettore di dati.
</Tip>
Successivamente, converti i datasets tokenizzati in TensorFlow datasets con il metodo [`to_tf_dataset`](https://huggingface.co/docs/datasets/package_reference/main_classes.html#datasets.Dataset.to_tf_dataset). Specifica il tuo input in `columns` e le tue etichette in `label_cols`:
```py
>>> tf_train_dataset = small_train_dataset.to_tf_dataset(
... columns=["attention_mask", "input_ids", "token_type_ids"],
... label_cols=["labels"],
... shuffle=True,
... collate_fn=data_collator,
... batch_size=8,
... )
>>> tf_validation_dataset = small_eval_dataset.to_tf_dataset(
... columns=["attention_mask", "input_ids", "token_type_ids"],
... label_cols=["labels"],
... shuffle=False,
... collate_fn=data_collator,
... batch_size=8,
... )
```
### Compilazione e addestramento
Carica un modello TensorFlow col numero atteso di etichette:
```py
>>> import tensorflow as tf
>>> from transformers import TFAutoModelForSequenceClassification
>>> model = TFAutoModelForSequenceClassification.from_pretrained("bert-base-cased", num_labels=5)
```
Poi compila e fai il fine-tuning del tuo modello usando [`fit`](https://keras.io/api/models/model_training_apis/) come faresti con qualsiasi altro modello di Keras:
```py
>>> model.compile(
... optimizer=tf.keras.optimizers.Adam(learning_rate=5e-5),
... loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
... metrics=tf.metrics.SparseCategoricalAccuracy(),
... )
>>> model.fit(tf_train_dataset, validation_data=tf_validation_dataset, epochs=3)
```
</tf>
</frameworkcontent>
<a id='pytorch_native'></a>
## Addestramento in PyTorch nativo
<frameworkcontent>
<pt>
<Youtube id="Dh9CL8fyG80"/>
[`Trainer`] si occupa del ciclo di addestramento e ti consente di mettere a punto un modello con una sola riga di codice. Per chi preferisse scrivere un proprio ciclo di addestramento personale, puoi anche fare il fine-tuning di un modello ๐ค Transformers in PyTorch nativo.
A questo punto, potresti avere bisogno di riavviare il tuo notebook o eseguire il seguente codice per liberare un po' di memoria:
```py
del model
del pytorch_model
del trainer
torch.cuda.empty_cache()
```
Successivamente, postprocessa manualmente il `tokenized_dataset` per prepararlo ad essere allenato.
1. Rimuovi la colonna `text` perchรฉ il modello non accetta testo grezzo come input:
```py
>>> tokenized_datasets = tokenized_datasets.remove_columns(["text"])
```
2. Rinomina la colonna `label` in `labels` perchรฉ il modello si aspetta che questo argomento si chiami `labels`:
```py
>>> tokenized_datasets = tokenized_datasets.rename_column("label", "labels")
```
3. Imposta il formato del dataset per farti restituire tensori di PyTorch all'interno delle liste:
```py
>>> tokenized_datasets.set_format("torch")
```
Poi crea un piccolo sottocampione del dataset come visto precedentemente per velocizzare il fine-tuning:
```py
>>> small_train_dataset = tokenized_datasets["train"].shuffle(seed=42).select(range(1000))
>>> small_eval_dataset = tokenized_datasets["test"].shuffle(seed=42).select(range(1000))
```
### DataLoader
Crea un `DataLoader` per i tuoi datasets di train e test cosรฌ puoi iterare sui lotti di dati:
```py
>>> from torch.utils.data import DataLoader
>>> train_dataloader = DataLoader(small_train_dataset, shuffle=True, batch_size=8)
>>> eval_dataloader = DataLoader(small_eval_dataset, batch_size=8)
```
Carica il tuo modello con il numero atteso di etichette:
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained("bert-base-cased", num_labels=5)
```
### Ottimizzatore e learning rate scheduler
Crea un ottimizzatore e il learning rate scheduler per fare il fine-tuning del modello. Usa l'ottimizzatore [`AdamW`](https://pytorch.org/docs/stable/generated/torch.optim.AdamW.html) di PyTorch:
```py
>>> from torch.optim import AdamW
>>> optimizer = AdamW(model.parameters(), lr=5e-5)
```
Crea il learning rate scheduler predefinito da [`Trainer`]:
```py
>>> from transformers import get_scheduler
>>> num_epochs = 3
>>> num_training_steps = num_epochs * len(train_dataloader)
>>> lr_scheduler = get_scheduler(
... name="linear", optimizer=optimizer, num_warmup_steps=0, num_training_steps=num_training_steps
... )
```
Infine specifica come `device` da usare una GPU se ne hai una. Altrimenti, l'addestramento su una CPU puรฒ richiedere diverse ore invece di un paio di minuti.
```py
>>> import torch
>>> device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
>>> model.to(device)
```
<Tip>
Ottieni l'accesso gratuito a una GPU sul cloud se non ne possiedi una usando un notebook sul web come [Colaboratory](https://colab.research.google.com/) o [SageMaker StudioLab](https://studiolab.sagemaker.aws/).
</Tip>
Ottimo, adesso possiamo addestrare! ๐ฅณ
### Training loop
Per tenere traccia dei tuoi progressi durante l'addestramento, usa la libreria [tqdm](https://tqdm.github.io/) per aggiungere una progress bar sopra il numero dei passi di addestramento:
```py
>>> from tqdm.auto import tqdm
>>> progress_bar = tqdm(range(num_training_steps))
>>> model.train()
>>> for epoch in range(num_epochs):
... for batch in train_dataloader:
... batch = {k: v.to(device) for k, v in batch.items()}
... outputs = model(**batch)
... loss = outputs.loss
... loss.backward()
... optimizer.step()
... lr_scheduler.step()
... optimizer.zero_grad()
... progress_bar.update(1)
```
### Metriche
Proprio come รจ necessario aggiungere una funzione di valutazione del [`Trainer`], รจ necessario fare lo stesso quando si scrive il proprio ciclo di addestramento. Ma invece di calcolare e riportare la metrica alla fine di ogni epoca, questa volta accumulerai tutti i batch con [`add_batch`](https://huggingface.co/docs/datasets/package_reference/main_classes?highlight=add_batch#datasets.Metric.add_batch) e calcolerai la metrica alla fine.
```py
>>> metric = load_metric("accuracy")
>>> model.eval()
>>> for batch in eval_dataloader:
... batch = {k: v.to(device) for k, v in batch.items()}
... with torch.no_grad():
... outputs = model(**batch)
... logits = outputs.logits
... predictions = torch.argmax(logits, dim=-1)
... metric.add_batch(predictions=predictions, references=batch["labels"])
>>> metric.compute()
```
</pt>
</frameworkcontent>
<a id='additional-resources'></a>
## Altre risorse
Per altri esempi sul fine-tuning, fai riferimento a:
- [๐ค Transformers Examples](https://github.com/huggingface/transformers/tree/main/examples) include scripts per addestrare compiti comuni di NLP in PyTorch e TensorFlow.
- [๐ค Transformers Notebooks](notebooks) contiene diversi notebooks su come mettere a punto un modello per compiti specifici in PyTorch e TensorFlow.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/index.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ๐ค Transformers
Machine Learning allo stato dell'arte per PyTorch, TensorFlow e JAX.
๐ค Transformers fornisce delle API per scaricare in modo semplice e allenare modelli pre-allenati allo stato dell'arte. L'utilizzo di modelli pre-allenati puรฒ ridurre i tuoi costi computazionali, l'impatto ambientale, e farti risparmiare il tempo che utilizzeresti per allenare un modello da zero. I modelli possono essere utilizzati in diverse modalitร come ad esempio:
* ๐ Testo: classificazione del testo, estrazione delle informazioni, rispondere a domande, riassumere, traduzione e generazione del testo in piรน di 100 lingue.
* ๐ผ๏ธ Immagini: classificazione di immagini, rilevazione di oggetti e segmentazione.
* ๐ฃ๏ธ Audio: riconoscimento vocale e classificazione dell'audio.
* ๐ Multimodale: rispondere a domande inerenti dati tabulari, riconoscimento ottico dei caratteri, estrazione di informazioni a partire da documenti scannerizzati, classificazione di video e risposta visuale a domande.
La nostra libreria supporta un'integrazione perfetta tra tre delle librerie per il deep learning piรน popolari: [PyTorch](https://pytorch.org/), [TensorFlow](https://www.tensorflow.org/) e [JAX](https://jax.readthedocs.io/en/latest/). Allena il tuo modello in tre righe di codice in un framework, e caricalo per l'inferenza in un altro.
Ogni architettura di ๐ค Transformers รจ definita in un modulo Python indipendente cosรฌ da poter essere personalizzata in modo semplice per la ricerca e gli esperimenti.
## Se stai cercando supporto personalizzato dal team di Hugging Face
<a target="_blank" href="https://huggingface.co/support">
<img alt="HuggingFace Expert Acceleration Program" src="https://huggingface.co/front/thumbnails/support.png" style="width: 100%; max-width: 600px; border: 1px solid #eee; border-radius: 4px; box-shadow: 0 1px 2px 0 rgba(0, 0, 0, 0.05);">
</a>
## Contenuti
La documentazione รจ organizzata in cinque parti:
- **INIZIARE** contiene un tour rapido e le istruzioni di installazione per cominciare ad utilizzare ๐ค Transformers.
- **TUTORIALS** รจ un buon posto da cui iniziare se per te la nostra libreria รจ nuova. Questa sezione ti aiuterร ad acquisire le competenze basilari di cui hai bisogno per iniziare ad utilizzare ๐ค Transformers.
- **GUIDE PRATICHE** ti mostrerร come raggiungere obiettivi specifici come fare fine-tuning di un modello pre-allenato per la modellizzazione del linguaggio o come creare una testa per un modello personalizzato.
- **GUIDE CONCETTUALI** fornisce discussioni e spiegazioni dei concetti sottostanti alle idee dietro ai modelli, compiti, e la filosofia di progettazione di ๐ค Transformers.
- **API** descrive ogni classe e funzione, raggruppate in:
- **CLASSI PRINCIPALI** per le classi principali che espongono le API importanti della libreria.
- **MODELLI** per le classi e le funzioni relative ad ogni modello implementato all'interno della libreria.
- **HELPERS INTERNI** per le classi e le funzioni che utilizziamo internamente.
La libreria attualmente contiene implementazioni in JAX, PyTorch e TensorFlow, pesi di modelli pre-allenati, script di utilizzo e strumenti di conversione per i seguenti modelli.
### Modelli supportati
<!--This list is updated automatically from the README with _make fix-copies_. Do not update manually! -->
1. **[ALBERT](model_doc/albert)** (da Google Research e l'Istituto Tecnologico di Chicago) rilasciato con il paper [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://arxiv.org/abs/1909.11942), da Zhenzhong Lan, Mingda Chen, Sebastian Goodman, Kevin Gimpel, Piyush Sharma, Radu Soricut.
1. **[ALIGN](model_doc/align)** (from Google Research) rilasciato con il paper [Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision](https://arxiv.org/abs/2102.05918) da Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc V. Le, Yunhsuan Sung, Zhen Li, Tom Duerig.
1. **[BART](model_doc/bart)** (da Facebook) rilasciato con il paper [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://arxiv.org/abs/1910.13461) da Mike Lewis, Yinhan Liu, Naman Goyal, Marjan Ghazvininejad, Abdelrahman Mohamed, Omer Levy, Ves Stoyanov e Luke Zettlemoyer.
1. **[BARThez](model_doc/barthez)** (da politecnico di รcole) rilasciato con il paper [BARThez: a Skilled Pretrained French Sequence-to-Sequence Model](https://arxiv.org/abs/2010.12321) da Moussa Kamal Eddine, Antoine J.-P. Tixier, Michalis Vazirgiannis.
1. **[BARTpho](model_doc/bartpho)** (da VinAI Research) rilasciato con il paper [BARTpho: Pre-trained Sequence-to-Sequence Models for Vietnamese](https://arxiv.org/abs/2109.09701) da Nguyen Luong Tran, Duong Minh Le e Dat Quoc Nguyen.
1. **[BEiT](model_doc/beit)** (da Microsoft) rilasciato con il paper [BEiT: BERT Pre-Training of Image Transformers](https://arxiv.org/abs/2106.08254) da Hangbo Bao, Li Dong, Furu Wei.
1. **[BERT](model_doc/bert)** (da Google) rilasciato con il paper [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) da Jacob Devlin, Ming-Wei Chang, Kenton Lee e Kristina Toutanova.
1. **[BERTweet](model_doc/bertweet)** (da VinAI Research) rilasciato con il paper [BERTweet: A pre-trained language model for English Tweets](https://aclanthology.org/2020.emnlp-demos.2/) da Dat Quoc Nguyen, Thanh Vu e Anh Tuan Nguyen.
1. **[BERT For Sequence Generation](model_doc/bert-generation)** (da Google) rilasciato con il paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) da Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
1. **[BigBird-RoBERTa](model_doc/big_bird)** (da Google Research) rilasciato con il paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) da Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
1. **[BigBird-Pegasus](model_doc/bigbird_pegasus)** (v Google Research) rilasciato con il paper [Big Bird: Transformers for Longer Sequences](https://arxiv.org/abs/2007.14062) da Manzil Zaheer, Guru Guruganesh, Avinava Dubey, Joshua Ainslie, Chris Alberti, Santiago Ontanon, Philip Pham, Anirudh Ravula, Qifan Wang, Li Yang, Amr Ahmed.
1. **[Blenderbot](model_doc/blenderbot)** (da Facebook) rilasciato con il paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) da Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
1. **[BlenderbotSmall](model_doc/blenderbot-small)** (da Facebook) rilasciato con il paper [Recipes for building an open-domain chatbot](https://arxiv.org/abs/2004.13637) da Stephen Roller, Emily Dinan, Naman Goyal, Da Ju, Mary Williamson, Yinhan Liu, Jing Xu, Myle Ott, Kurt Shuster, Eric M. Smith, Y-Lan Boureau, Jason Weston.
1. **[BORT](model_doc/bort)** (da Alexa) rilasciato con il paper [Optimal Subarchitecture Extraction For BERT](https://arxiv.org/abs/2010.10499) da Adrian de Wynter e Daniel J. Perry.
1. **[ByT5](model_doc/byt5)** (da Google Research) rilasciato con il paper [ByT5: Towards a token-free future with pre-trained byte-to-byte models](https://arxiv.org/abs/2105.13626) da Linting Xue, Aditya Barua, Noah Constant, Rami Al-Rfou, Sharan Narang, Mihir Kale, Adam Roberts, Colin Raffel.
1. **[CamemBERT](model_doc/camembert)** (da Inria/Facebook/Sorbonne) rilasciato con il paper [CamemBERT: a Tasty French Language Model](https://arxiv.org/abs/1911.03894) da Louis Martin*, Benjamin Muller*, Pedro Javier Ortiz Suรกrez*, Yoann Dupont, Laurent Romary, รric Villemonte de la Clergerie, Djamรฉ Seddah e Benoรฎt Sagot.
1. **[CANINE](model_doc/canine)** (da Google Research) rilasciato con il paper [CANINE: Pre-training an Efficient Tokenization-Free Encoder for Language Representation](https://arxiv.org/abs/2103.06874) da Jonathan H. Clark, Dan Garrette, Iulia Turc, John Wieting.
1. **[ConvNeXT](model_doc/convnext)** (da Facebook AI) rilasciato con il paper [A ConvNet for the 2020s](https://arxiv.org/abs/2201.03545) da Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, Saining Xie.
1. **[ConvNeXTV2](model_doc/convnextv2)** (da Facebook AI) rilasciato con il paper [ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders](https://arxiv.org/abs/2301.00808) da Sanghyun Woo, Shoubhik Debnath, Ronghang Hu, Xinlei Chen, Zhuang Liu, In So Kweon, Saining Xie.
1. **[CLIP](model_doc/clip)** (da OpenAI) rilasciato con il paper [Learning Transferable Visual Models From Natural Language Supervision](https://arxiv.org/abs/2103.00020) da Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, Ilya Sutskever.
1. **[ConvBERT](model_doc/convbert)** (da YituTech) rilasciato con il paper [ConvBERT: Improving BERT with Span-based Dynamic Convolution](https://arxiv.org/abs/2008.02496) da Zihang Jiang, Weihao Yu, Daquan Zhou, Yunpeng Chen, Jiashi Feng, Shuicheng Yan.
1. **[CPM](model_doc/cpm)** (dalla Universitร di Tsinghua) rilasciato con il paper [CPM: A Large-scale Generative Chinese Pre-trained Language Model](https://arxiv.org/abs/2012.00413) da Zhengyan Zhang, Xu Han, Hao Zhou, Pei Ke, Yuxian Gu, Deming Ye, Yujia Qin, Yusheng Su, Haozhe Ji, Jian Guan, Fanchao Qi, Xiaozhi Wang, Yanan Zheng, Guoyang Zeng, Huanqi Cao, Shengqi Chen, Daixuan Li, Zhenbo Sun, Zhiyuan Liu, Minlie Huang, Wentao Han, Jie Tang, Juanzi Li, Xiaoyan Zhu, Maosong Sun.
1. **[CTRL](model_doc/ctrl)** (da Salesforce) rilasciato con il paper [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://arxiv.org/abs/1909.05858) da Nitish Shirish Keskar*, Bryan McCann*, Lav R. Varshney, Caiming Xiong e Richard Socher.
1. **[CvT](model_doc/cvt)** (da Microsoft) rilasciato con il paper [CvT: Introducing Convolutions to Vision Transformers](https://arxiv.org/abs/2103.15808) da Haiping Wu, Bin Xiao, Noel Codella, Mengchen Liu, Xiyang Dai, Lu Yuan, Lei Zhang.
1. **[Data2Vec](model_doc/data2vec)** (da Facebook) rilasciato con il paper [Data2Vec: A General Framework for Self-supervised Learning in Speech, Vision and Language](https://arxiv.org/abs/2202.03555) da Alexei Baevski, Wei-Ning Hsu, Qiantong Xu, Arun Babu, Jiatao Gu, Michael Auli.
1. **[DeBERTa](model_doc/deberta)** (da Microsoft) rilasciato con il paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) da Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
1. **[DeBERTa-v2](model_doc/deberta-v2)** (da Microsoft) rilasciato con il paper [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://arxiv.org/abs/2006.03654) da Pengcheng He, Xiaodong Liu, Jianfeng Gao, Weizhu Chen.
1. **[Decision Transformer](model_doc/decision_transformer)** (da Berkeley/Facebook/Google) rilasciato con il paper [Decision Transformer: Reinforcement Learning via Sequence Modeling](https://arxiv.org/abs/2106.01345) da Lili Chen, Kevin Lu, Aravind Rajeswaran, Kimin Lee, Aditya Grover, Michael Laskin, Pieter Abbeel, Aravind Srinivas, Igor Mordatch.
1. **[DiT](model_doc/dit)** (da Microsoft Research) rilasciato con il paper [DiT: Self-supervised Pre-training for Document Image Transformer](https://arxiv.org/abs/2203.02378) da Junlong Li, Yiheng Xu, Tengchao Lv, Lei Cui, Cha Zhang, Furu Wei.
1. **[DeiT](model_doc/deit)** (da Facebook) rilasciato con il paper [Training data-efficient image transformers & distillation through attention](https://arxiv.org/abs/2012.12877) da Hugo Touvron, Matthieu Cord, Matthijs Douze, Francisco Massa, Alexandre Sablayrolles, Hervรฉ Jรฉgou.
1. **[DETR](model_doc/detr)** (da Facebook) rilasciato con il paper [End-to-End Object Detection with Transformers](https://arxiv.org/abs/2005.12872) da Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, Sergey Zagoruyko.
1. **[DialoGPT](model_doc/dialogpt)** (da Microsoft Research) rilasciato con il paper [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://arxiv.org/abs/1911.00536) da Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, Bill Dolan.
1. **[DistilBERT](model_doc/distilbert)** (da HuggingFace), rilasciato assieme al paper [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://arxiv.org/abs/1910.01108) da Victor Sanh, Lysandre Debut e Thomas Wolf. La stessa tecnica รจ stata applicata per comprimere GPT2 in [DistilGPT2](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), RoBERTa in [DistilRoBERTa](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation), Multilingual BERT in [DistilmBERT](https://github.com/huggingface/transformers/tree/main/examples/research_projects/distillation) and a German version of DistilBERT.
1. **[DPR](model_doc/dpr)** (da Facebook) rilasciato con il paper [Dense Passage Retrieval for Open-Domain Question Answering](https://arxiv.org/abs/2004.04906) da Vladimir Karpukhin, Barlas Oฤuz, Sewon Min, Patrick Lewis, Ledell Wu, Sergey Edunov, Danqi Chen, e Wen-tau Yih.
1. **[DPT](master/model_doc/dpt)** (da Intel Labs) rilasciato con il paper [Vision Transformers for Dense Prediction](https://arxiv.org/abs/2103.13413) da Renรฉ Ranftl, Alexey Bochkovskiy, Vladlen Koltun.
1. **[EfficientNet](model_doc/efficientnet)** (from Google Research) released with the paper [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://arxiv.org/abs/1905.11946) by Mingxing Tan and Quoc V. Le.
1. **[EncoderDecoder](model_doc/encoder-decoder)** (da Google Research) rilasciato con il paper [Leveraging Pre-trained Checkpoints for Sequence Generation Tasks](https://arxiv.org/abs/1907.12461) da Sascha Rothe, Shashi Narayan, Aliaksei Severyn.
1. **[ELECTRA](model_doc/electra)** (da Google Research/Stanford University) rilasciato con il paper [ELECTRA: Pre-training text encoders as discriminators rather than generators](https://arxiv.org/abs/2003.10555) da Kevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. Manning.
1. **[FlauBERT](model_doc/flaubert)** (da CNRS) rilasciato con il paper [FlauBERT: Unsupervised Language Model Pre-training for French](https://arxiv.org/abs/1912.05372) da Hang Le, Loรฏc Vial, Jibril Frej, Vincent Segonne, Maximin Coavoux, Benjamin Lecouteux, Alexandre Allauzen, Benoรฎt Crabbรฉ, Laurent Besacier, Didier Schwab.
1. **[FLAVA](model_doc/flava)** (da Facebook AI) rilasciato con il paper [FLAVA: A Foundational Language And Vision Alignment Model](https://arxiv.org/abs/2112.04482) da Amanpreet Singh, Ronghang Hu, Vedanuj Goswami, Guillaume Couairon, Wojciech Galuba, Marcus Rohrbach, e Douwe Kiela.
1. **[FNet](model_doc/fnet)** (da Google Research) rilasciato con il paper [FNet: Mixing Tokens with Fourier Transforms](https://arxiv.org/abs/2105.03824) da James Lee-Thorp, Joshua Ainslie, Ilya Eckstein, Santiago Ontanon.
1. **[Funnel Transformer](model_doc/funnel)** (da CMU/Google Brain) rilasciato con il paper [Funnel-Transformer: Filtering out Sequential Redundancy for Efficient Language Processing](https://arxiv.org/abs/2006.03236) da Zihang Dai, Guokun Lai, Yiming Yang, Quoc V. Le.
1. **[GLPN](model_doc/glpn)** (da KAIST) rilasciato con il paper [Global-Local Path Networks for Monocular Depth Estimation with Vertical CutDepth](https://arxiv.org/abs/2201.07436) da Doyeon Kim, Woonghyun Ga, Pyungwhan Ahn, Donggyu Joo, Sehwan Chun, Junmo Kim.
1. **[GPT](model_doc/openai-gpt)** (da OpenAI) rilasciato con il paper [Improving Language Understanding by Generative Pre-Training](https://blog.openai.com/language-unsupervised/) da Alec Radford, Karthik Narasimhan, Tim Salimans e Ilya Sutskever.
1. **[GPT-2](model_doc/gpt2)** (da OpenAI) rilasciato con il paper [Language Models are Unsupervised Multitask Learners](https://blog.openai.com/better-language-models/) da Alec Radford*, Jeffrey Wu*, Rewon Child, David Luan, Dario Amodei** e Ilya Sutskever**.
1. **[GPT-J](model_doc/gptj)** (da EleutherAI) rilasciato nel repository [kingoflolz/mesh-transformer-jax](https://github.com/kingoflolz/mesh-transformer-jax/) da Ben Wang e Aran Komatsuzaki.
1. **[GPT Neo](model_doc/gpt_neo)** (da EleutherAI) rilasciato nel repository [EleutherAI/gpt-neo](https://github.com/EleutherAI/gpt-neo) da Sid Black, Stella Biderman, Leo Gao, Phil Wang e Connor Leahy.
1. **[GPT NeoX](model_doc/gpt_neox)** (da EleutherAI) rilasciato con il paper [GPT-NeoX-20B: An Open-Source Autoregressive Language Model](https://arxiv.org/abs/2204.06745) da Sid Black, Stella Biderman, Eric Hallahan, Quentin Anthony, Leo Gao, Laurence Golding, Horace He, Connor Leahy, Kyle McDonell, Jason Phang, Michael Pieler, USVSN Sai Prashanth, Shivanshu Purohit, Laria Reynolds, Jonathan Tow, Ben Wang, Samuel Weinbach
1. **[Hubert](model_doc/hubert)** (da Facebook) rilasciato con il paper [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://arxiv.org/abs/2106.07447) da Wei-Ning Hsu, Benjamin Bolte, Yao-Hung Hubert Tsai, Kushal Lakhotia, Ruslan Salakhutdinov, Abdelrahman Mohamed.
1. **[I-BERT](model_doc/ibert)** (da Berkeley) rilasciato con il paper [I-BERT: Integer-only BERT Quantization](https://arxiv.org/abs/2101.01321) da Sehoon Kim, Amir Gholami, Zhewei Yao, Michael W. Mahoney, Kurt Keutzer.
1. **[ImageGPT](model_doc/imagegpt)** (da OpenAI) rilasciato con il paper [Generative Pretraining from Pixels](https://openai.com/blog/image-gpt/) da Mark Chen, Alec Radford, Rewon Child, Jeffrey Wu, Heewoo Jun, David Luan, Ilya Sutskever.
1. **[LayoutLM](model_doc/layoutlm)** (da Microsoft Research Asia) rilasciato con il paper [LayoutLM: Pre-training of Text and Layout for Document Image Understanding](https://arxiv.org/abs/1912.13318) da Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou.
1. **[LayoutLMv2](model_doc/layoutlmv2)** (da Microsoft Research Asia) rilasciato con il paper [LayoutLMv2: Multi-modal Pre-training for Visually-Rich Document Understanding](https://arxiv.org/abs/2012.14740) da Yang Xu, Yiheng Xu, Tengchao Lv, Lei Cui, Furu Wei, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Wanxiang Che, Min Zhang, Lidong Zhou.
1. **[LayoutLMv3](model_doc/layoutlmv3)** (da Microsoft Research Asia) rilasciato con il paper [LayoutLMv3: Pre-training for Document AI with Unified Text and Image Masking](https://arxiv.org/abs/2204.08387) da Yupan Huang, Tengchao Lv, Lei Cui, Yutong Lu, Furu Wei.
1. **[LayoutXLM](model_doc/layoutlxlm)** (da Microsoft Research Asia) rilasciato con il paper [LayoutXLM: Multimodal Pre-training for Multilingual Visually-rich Document Understanding](https://arxiv.org/abs/2104.08836) da Yiheng Xu, Tengchao Lv, Lei Cui, Guoxin Wang, Yijuan Lu, Dinei Florencio, Cha Zhang, Furu Wei.
1. **[LED](model_doc/led)** (da AllenAI) rilasciato con il paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) da Iz Beltagy, Matthew E. Peters, Arman Cohan.
1. **[Longformer](model_doc/longformer)** (da AllenAI) rilasciato con il paper [Longformer: The Long-Document Transformer](https://arxiv.org/abs/2004.05150) da Iz Beltagy, Matthew E. Peters, Arman Cohan.
1. **[LUKE](model_doc/luke)** (da Studio Ousia) rilasciato con il paper [LUKE: Deep Contextualized Entity Representations with Entity-aware Self-attention](https://arxiv.org/abs/2010.01057) da Ikuya Yamada, Akari Asai, Hiroyuki Shindo, Hideaki Takeda, Yuji Matsumoto.
1. **[mLUKE](model_doc/mluke)** (da Studio Ousia) rilasciato con il paper [mLUKE: The Power of Entity Representations in Multilingual Pretrained Language Models](https://arxiv.org/abs/2110.08151) da Ryokan Ri, Ikuya Yamada, e Yoshimasa Tsuruoka.
1. **[LXMERT](model_doc/lxmert)** (da UNC Chapel Hill) rilasciato con il paper [LXMERT: Learning Cross-Modality Encoder Representations from Transformers for Open-Domain Question Answering](https://arxiv.org/abs/1908.07490) da Hao Tan e Mohit Bansal.
1. **[M2M100](model_doc/m2m_100)** (da Facebook) rilasciato con il paper [Beyond English-Centric Multilingual Machine Translation](https://arxiv.org/abs/2010.11125) da Angela Fan, Shruti Bhosale, Holger Schwenk, Zhiyi Ma, Ahmed El-Kishky, Siddharth Goyal, Mandeep Baines, Onur Celebi, Guillaume Wenzek, Vishrav Chaudhary, Naman Goyal, Tom Birch, Vitaliy Liptchinsky, Sergey Edunov, Edouard Grave, Michael Auli, Armand Joulin.
1. **[MarianMT](model_doc/marian)** Modello di machine learning per le traduzioni allenato utilizzando i dati [OPUS](http://opus.nlpl.eu/) di Jรถrg Tiedemann. Il [Framework Marian](https://marian-nmt.github.io/) รจ stato sviluppato dal Microsoft Translator Team.
1. **[Mask2Former](model_doc/mask2former)** (da FAIR e UIUC) rilasciato con il paper [Masked-attention Mask Transformer for Universal Image Segmentation](https://arxiv.org/abs/2112.01527) da Bowen Cheng, Ishan Misra, Alexander G. Schwing, Alexander Kirillov, Rohit Girdhar.
1. **[MaskFormer](model_doc/maskformer)** (da Meta e UIUC) rilasciato con il paper [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://arxiv.org/abs/2107.06278) da Bowen Cheng, Alexander G. Schwing, Alexander Kirillov.
1. **[MBart](model_doc/mbart)** (da Facebook) rilasciato con il paper [Multilingual Denoising Pre-training for Neural Machine Translation](https://arxiv.org/abs/2001.08210) da Yinhan Liu, Jiatao Gu, Naman Goyal, Xian Li, Sergey Edunov, Marjan Ghazvininejad, Mike Lewis, Luke Zettlemoyer.
1. **[MBart-50](model_doc/mbart)** (da Facebook) rilasciato con il paper [Multilingual Translation with Extensible Multilingual Pretraining and Finetuning](https://arxiv.org/abs/2008.00401) da Yuqing Tang, Chau Tran, Xian Li, Peng-Jen Chen, Naman Goyal, Vishrav Chaudhary, Jiatao Gu, Angela Fan.
1. **[Megatron-BERT](model_doc/megatron-bert)** (da NVIDIA) rilasciato con il paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) da Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper e Bryan Catanzaro.
1. **[Megatron-GPT2](model_doc/megatron_gpt2)** (da NVIDIA) rilasciato con il paper [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://arxiv.org/abs/1909.08053) da Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper e Bryan Catanzaro.
1. **[MPNet](model_doc/mpnet)** (da Microsoft Research) rilasciato con il paper [MPNet: Masked and Permuted Pre-training for Language Understanding](https://arxiv.org/abs/2004.09297) da Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, Tie-Yan Liu.
1. **[MT5](model_doc/mt5)** (da Google AI) rilasciato con il paper [mT5: A massively multilingual pre-trained text-to-text transformer](https://arxiv.org/abs/2010.11934) da Linting Xue, Noah Constant, Adam Roberts, Mihir Kale, Rami Al-Rfou, Aditya Siddhant, Aditya Barua, Colin Raffel.
1. **[Nystrรถmformer](model_doc/nystromformer)** (dalla Universitร del Wisconsin - Madison) rilasciato con il paper [Nystrรถmformer: A Nystrรถm-Based Algorithm for Approximating Self-Attention](https://arxiv.org/abs/2102.03902) da Yunyang Xiong, Zhanpeng Zeng, Rudrasis Chakraborty, Mingxing Tan, Glenn Fung, Yin Li, Vikas Singh.
1. **[OneFormer](model_doc/oneformer)** (da SHI Labs) rilasciato con il paper [OneFormer: One Transformer to Rule Universal Image Segmentation](https://arxiv.org/abs/2211.06220) da Jitesh Jain, Jiachen Li, MangTik Chiu, Ali Hassani, Nikita Orlov, Humphrey Shi.
1. **[OPT](master/model_doc/opt)** (da Meta AI) rilasciato con il paper [OPT: Open Pre-trained Transformer Language Models](https://arxiv.org/abs/2205.01068) da Susan Zhang, Stephen Roller, Naman Goyal, Mikel Artetxe, Moya Chen, Shuohui Chen et al.
1. **[Pegasus](model_doc/pegasus)** (da Google) rilasciato con il paper [PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization](https://arxiv.org/abs/1912.08777) da Jingqing Zhang, Yao Zhao, Mohammad Saleh e Peter J. Liu.
1. **[Perceiver IO](model_doc/perceiver)** (da Deepmind) rilasciato con il paper [Perceiver IO: A General Architecture for Structured Inputs & Outputs](https://arxiv.org/abs/2107.14795) da Andrew Jaegle, Sebastian Borgeaud, Jean-Baptiste Alayrac, Carl Doersch, Catalin Ionescu, David Ding, Skanda Koppula, Daniel Zoran, Andrew Brock, Evan Shelhamer, Olivier Hรฉnaff, Matthew M. Botvinick, Andrew Zisserman, Oriol Vinyals, Joรฃo Carreira.
1. **[PhoBERT](model_doc/phobert)** (da VinAI Research) rilasciato con il paper [PhoBERT: Pre-trained language models for Vietnamese](https://www.aclweb.org/anthology/2020.findings-emnlp.92/) da Dat Quoc Nguyen e Anh Tuan Nguyen.
1. **[PLBart](model_doc/plbart)** (da UCLA NLP) rilasciato con il paper [Unified Pre-training for Program Understanding and Generation](https://arxiv.org/abs/2103.06333) da Wasi Uddin Ahmad, Saikat Chakraborty, Baishakhi Ray, Kai-Wei Chang.
1. **[PoolFormer](model_doc/poolformer)** (da Sea AI Labs) rilasciato con il paper [MetaFormer is Actually What You Need for Vision](https://arxiv.org/abs/2111.11418) da Yu, Weihao e Luo, Mi e Zhou, Pan e Si, Chenyang e Zhou, Yichen e Wang, Xinchao e Feng, Jiashi e Yan, Shuicheng.
1. **[ProphetNet](model_doc/prophetnet)** (da Microsoft Research) rilasciato con il paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) da Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang e Ming Zhou.
1. **[QDQBert](model_doc/qdqbert)** (da NVIDIA) rilasciato con il paper [Integer Quantization for Deep Learning Inference: Principles and Empirical Evaluation](https://arxiv.org/abs/2004.09602) da Hao Wu, Patrick Judd, Xiaojie Zhang, Mikhail Isaev e Paulius Micikevicius.
1. **[REALM](model_doc/realm.html)** (da Google Research) rilasciato con il paper [REALM: Retrieval-Augmented Language Model Pre-Training](https://arxiv.org/abs/2002.08909) da Kelvin Guu, Kenton Lee, Zora Tung, Panupong Pasupat e Ming-Wei Chang.
1. **[Reformer](model_doc/reformer)** (da Google Research) rilasciato con il paper [Reformer: The Efficient Transformer](https://arxiv.org/abs/2001.04451) da Nikita Kitaev, ลukasz Kaiser, Anselm Levskaya.
1. **[RemBERT](model_doc/rembert)** (da Google Research) rilasciato con il paper [Rethinking embedding coupling in pre-trained language models](https://arxiv.org/abs/2010.12821) da Hyung Won Chung, Thibault Fรฉvry, Henry Tsai, M. Johnson, Sebastian Ruder.
1. **[RegNet](model_doc/regnet)** (da META Platforms) rilasciato con il paper [Designing Network Design Space](https://arxiv.org/abs/2003.13678) da Ilija Radosavovic, Raj Prateek Kosaraju, Ross Girshick, Kaiming He, Piotr Dollรกr.
1. **[ResNet](model_doc/resnet)** (da Microsoft Research) rilasciato con il paper [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385) da Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun.
1. **[RoBERTa](model_doc/roberta)** (da Facebook), rilasciato assieme al paper [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://arxiv.org/abs/1907.11692) da Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, Veselin Stoyanov.
1. **[RoFormer](model_doc/roformer)** (da ZhuiyiTechnology), rilasciato assieme al paper [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://arxiv.org/abs/2104.09864) da Jianlin Su e Yu Lu e Shengfeng Pan e Bo Wen e Yunfeng Liu.
1. **[SegFormer](model_doc/segformer)** (da NVIDIA) rilasciato con il paper [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://arxiv.org/abs/2105.15203) da Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M. Alvarez, Ping Luo.
1. **[SEW](model_doc/sew)** (da ASAPP) rilasciato con il paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) da Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
1. **[SEW-D](model_doc/sew_d)** (da ASAPP) rilasciato con il paper [Performance-Efficiency Trade-offs in Unsupervised Pre-training for Speech Recognition](https://arxiv.org/abs/2109.06870) da Felix Wu, Kwangyoun Kim, Jing Pan, Kyu Han, Kilian Q. Weinberger, Yoav Artzi.
1. **[SpeechToTextTransformer](model_doc/speech_to_text)** (da Facebook), rilasciato assieme al paper [fairseq S2T: Fast Speech-to-Text Modeling with fairseq](https://arxiv.org/abs/2010.05171) da Changhan Wang, Yun Tang, Xutai Ma, Anne Wu, Dmytro Okhonko, Juan Pino.
1. **[SpeechToTextTransformer2](model_doc/speech_to_text_2)** (da Facebook), rilasciato assieme al paper [Large-Scale Self- and Semi-Supervised Learning for Speech Translation](https://arxiv.org/abs/2104.06678) da Changhan Wang, Anne Wu, Juan Pino, Alexei Baevski, Michael Auli, Alexis Conneau.
1. **[Splinter](model_doc/splinter)** (dalla Universitร di Tel Aviv), rilasciato assieme al paper [Few-Shot Question Answering by Pretraining Span Selection](https://arxiv.org/abs/2101.00438) da Ori Ram, Yuval Kirstain, Jonathan Berant, Amir Globerson, Omer Levy.
1. **[SqueezeBert](model_doc/squeezebert)** (da Berkeley) rilasciato con il paper [SqueezeBERT: What can computer vision teach NLP about efficient neural networks?](https://arxiv.org/abs/2006.11316) da Forrest N. Iandola, Albert E. Shaw, Ravi Krishna, e Kurt W. Keutzer.
1. **[Swin Transformer](model_doc/swin)** (da Microsoft) rilasciato con il paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) da Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, Baining Guo.
1. **[T5](model_doc/t5)** (da Google AI) rilasciato con il paper [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://arxiv.org/abs/1910.10683) da Colin Raffel e Noam Shazeer e Adam Roberts e Katherine Lee e Sharan Narang e Michael Matena e Yanqi Zhou e Wei Li e Peter J. Liu.
1. **[T5v1.1](model_doc/t5v1.1)** (da Google AI) rilasciato nel repository [google-research/text-to-text-transfer-transformer](https://github.com/google-research/text-to-text-transfer-transformer/blob/main/released_checkpoints.md#t511) da Colin Raffel e Noam Shazeer e Adam Roberts e Katherine Lee e Sharan Narang e Michael Matena e Yanqi Zhou e Wei Li e Peter J. Liu.
1. **[TAPAS](model_doc/tapas)** (da Google AI) rilasciato con il paper [TAPAS: Weakly Supervised Table Parsing via Pre-training](https://arxiv.org/abs/2004.02349) da Jonathan Herzig, Paweล Krzysztof Nowak, Thomas Mรผller, Francesco Piccinno e Julian Martin Eisenschlos.
1. **[TAPEX](model_doc/tapex)** (da Microsoft Research) rilasciato con il paper [TAPEX: Table Pre-training via Learning a Neural SQL Executor](https://arxiv.org/abs/2107.07653) da Qian Liu, Bei Chen, Jiaqi Guo, Morteza Ziyadi, Zeqi Lin, Weizhu Chen, Jian-Guang Lou.
1. **[Trajectory Transformer](model_doc/trajectory_transformers)** (dall'Universitร della California a Berkeley) rilasciato con il paper [Offline Reinforcement Learning as One Big Sequence Modeling Problem](https://arxiv.org/abs/2106.02039) da Michael Janner, Qiyang Li, Sergey Levine
1. **[Transformer-XL](model_doc/transfo-xl)** (da Google/CMU) rilasciato con il paper [Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context](https://arxiv.org/abs/1901.02860) da Zihang Dai*, Zhilin Yang*, Yiming Yang, Jaime Carbonell, Quoc V. Le, Ruslan Salakhutdinov.
1. **[TrOCR](model_doc/trocr)** (da Microsoft), rilasciato assieme al paper [TrOCR: Transformer-based Optical Character Recognition with Pre-trained Models](https://arxiv.org/abs/2109.10282) da Minghao Li, Tengchao Lv, Lei Cui, Yijuan Lu, Dinei Florencio, Cha Zhang, Zhoujun Li, Furu Wei.
1. **[UniSpeech](model_doc/unispeech)** (da Microsoft Research) rilasciato con il paper [UniSpeech: Unified Speech Representation Learning with Labeled and Unlabeled Data](https://arxiv.org/abs/2101.07597) da Chengyi Wang, Yu Wu, Yao Qian, Kenichi Kumatani, Shujie Liu, Furu Wei, Michael Zeng, Xuedong Huang.
1. **[UniSpeechSat](model_doc/unispeech-sat)** (da Microsoft Research) rilasciato con il paper [UNISPEECH-SAT: UNIVERSAL SPEECH REPRESENTATION LEARNING WITH SPEAKER AWARE PRE-TRAINING](https://arxiv.org/abs/2110.05752) da Sanyuan Chen, Yu Wu, Chengyi Wang, Zhengyang Chen, Zhuo Chen, Shujie Liu, Jian Wu, Yao Qian, Furu Wei, Jinyu Li, Xiangzhan Yu.
1. **[VAN](model_doc/van)** (dalle Universitร di Tsinghua e Nankai) rilasciato con il paper [Visual Attention Network](https://arxiv.org/abs/2202.09741) da Meng-Hao Guo, Cheng-Ze Lu, Zheng-Ning Liu, Ming-Ming Cheng, Shi-Min Hu.
1. **[ViLT](model_doc/vilt)** (da NAVER AI Lab/Kakao Enterprise/Kakao Brain) rilasciato con il paper [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://arxiv.org/abs/2102.03334) da Wonjae Kim, Bokyung Son, Ildoo Kim.
1. **[Vision Transformer (ViT)](model_doc/vit)** (da Google AI) rilasciato con il paper [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/abs/2010.11929) da Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Sylvain Gelly, Jakob Uszkoreit, Neil Houlsby.
1. **[ViTMAE](model_doc/vit_mae)** (da Meta AI) rilasciato con il paper [Masked Autoencoders Are Scalable Vision Learners](https://arxiv.org/abs/2111.06377) da Kaiming He, Xinlei Chen, Saining Xie, Yanghao Li, Piotr Dollรกr, Ross Girshick.
1. **[VisualBERT](model_doc/visual_bert)** (da UCLA NLP) rilasciato con il paper [VisualBERT: A Simple and Performant Baseline for Vision and Language](https://arxiv.org/pdf/1908.03557) da Liunian Harold Li, Mark Yatskar, Da Yin, Cho-Jui Hsieh, Kai-Wei Chang.
1. **[WavLM](model_doc/wavlm)** (da Microsoft Research) rilasciato con il paper [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) da Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei.
1. **[Wav2Vec2](model_doc/wav2vec2)** (da Facebook AI) rilasciato con il paper [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://arxiv.org/abs/2006.11477) da Alexei Baevski, Henry Zhou, Abdelrahman Mohamed, Michael Auli.
1. **[Wav2Vec2Phoneme](model_doc/wav2vec2_phoneme)** (da Facebook AI) rilasciato con il paper [Simple and Effective Zero-shot Cross-lingual Phoneme Recognition](https://arxiv.org/abs/2109.11680) da Qiantong Xu, Alexei Baevski, Michael Auli.
1. **[XGLM](model_doc/xglm)** (da Facebook AI) rilasciato con il paper [Few-shot Learning with Multilingual Language Models](https://arxiv.org/abs/2112.10668) da Xi Victoria Lin, Todor Mihaylov, Mikel Artetxe, Tianlu Wang, Shuohui Chen, Daniel Simig, Myle Ott, Naman Goyal, Shruti Bhosale, Jingfei Du, Ramakanth Pasunuru, Sam Shleifer, Punit Singh Koura, Vishrav Chaudhary, Brian O'Horo, Jeff Wang, Luke Zettlemoyer, Zornitsa Kozareva, Mona Diab, Veselin Stoyanov, Xian Li.
1. **[XLM](model_doc/xlm)** (v Facebook) rilasciato assieme al paper [Cross-lingual Language Model Pretraining](https://arxiv.org/abs/1901.07291) da Guillaume Lample e Alexis Conneau.
1. **[XLM-ProphetNet](model_doc/xlm-prophetnet)** (da Microsoft Research) rilasciato con il paper [ProphetNet: Predicting Future N-gram for Sequence-to-Sequence Pre-training](https://arxiv.org/abs/2001.04063) da Yu Yan, Weizhen Qi, Yeyun Gong, Dayiheng Liu, Nan Duan, Jiusheng Chen, Ruofei Zhang e Ming Zhou.
1. **[XLM-RoBERTa](model_doc/xlm-roberta)** (da Facebook AI), rilasciato assieme al paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) da Alexis Conneau*, Kartikay Khandelwal*, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmรกn, Edouard Grave, Myle Ott, Luke Zettlemoyer e Veselin Stoyanov.
1. **[XLM-RoBERTa-XL](model_doc/xlm-roberta-xl)** (da Facebook AI), rilasciato assieme al paper [Larger-Scale Transformers for Multilingual Masked Language Modeling](https://arxiv.org/abs/2105.00572) da Naman Goyal, Jingfei Du, Myle Ott, Giri Anantharaman, Alexis Conneau.
1. **[XLNet](model_doc/xlnet)** (da Google/CMU) rilasciato con il paper [โXLNet: Generalized Autoregressive Pretraining for Language Understanding](https://arxiv.org/abs/1906.08237) da Zhilin Yang*, Zihang Dai*, Yiming Yang, Jaime Carbonell, Ruslan Salakhutdinov, Quoc V. Le.
1. **[XLSR-Wav2Vec2](model_doc/xlsr_wav2vec2)** (da Facebook AI) rilasciato con il paper [Unsupervised Cross-Lingual Representation Learning For Speech Recognition](https://arxiv.org/abs/2006.13979) da Alexis Conneau, Alexei Baevski, Ronan Collobert, Abdelrahman Mohamed, Michael Auli.
1. **[XLS-R](model_doc/xls_r)** (da Facebook AI) rilasciato con il paper [XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale](https://arxiv.org/abs/2111.09296) da Arun Babu, Changhan Wang, Andros Tjandra, Kushal Lakhotia, Qiantong Xu, Naman Goyal, Kritika Singh, Patrick von Platen, Yatharth Saraf, Juan Pino, Alexei Baevski, Alexis Conneau, Michael Auli.
1. **[YOLOS](model_doc/yolos)** (dalla Universitร della scienza e tecnologia di Huazhong) rilasciato con il paper [You Only Look at One Sequence: Rethinking Transformer in Vision through Object Detection](https://arxiv.org/abs/2106.00666) da Yuxin Fang, Bencheng Liao, Xinggang Wang, Jiemin Fang, Jiyang Qi, Rui Wu, Jianwei Niu, Wenyu Liu.
1. **[YOSO](model_doc/yoso)** (dall'Universitร del Wisconsin - Madison) rilasciato con il paper [You Only Sample (Almost) Once: Linear Cost Self-Attention Via Bernoulli Sampling](https://arxiv.org/abs/2111.09714) da Zhanpeng Zeng, Yunyang Xiong, Sathya N. Ravi, Shailesh Acharya, Glenn Fung, Vikas Singh.
### Framework supportati
La tabella seguente rappresenta il supporto attuale nella libreria per ognuno di questi modelli, si puรฒ identificare se questi hanno un Python
tokenizer (chiamato "slow"). Un tokenizer "fast" supportato dalla libreria ๐ค Tokenizers, e se hanno supporto in Jax (via Flax), PyTorch, e/o TensorFlow.
<!--This table is updated automatically from the auto modules with _make fix-copies_. Do not update manually!-->
| Model | Tokenizer slow | Tokenizer fast | PyTorch support | TensorFlow support | Flax Support |
|:---------------------------:|:--------------:|:--------------:|:---------------:|:------------------:|:------------:|
| ALBERT | โ
| โ
| โ
| โ
| โ
|
| BART | โ
| โ
| โ
| โ
| โ
|
| BEiT | โ | โ | โ
| โ | โ
|
| BERT | โ
| โ
| โ
| โ
| โ
|
| Bert Generation | โ
| โ | โ
| โ | โ |
| BigBird | โ
| โ
| โ
| โ | โ
|
| BigBirdPegasus | โ | โ | โ
| โ | โ |
| Blenderbot | โ
| โ
| โ
| โ
| โ
|
| BlenderbotSmall | โ
| โ
| โ
| โ
| โ
|
| CamemBERT | โ
| โ
| โ
| โ
| โ |
| Canine | โ
| โ | โ
| โ | โ |
| CLIP | โ
| โ
| โ
| โ
| โ
|
| ConvBERT | โ
| โ
| โ
| โ
| โ |
| ConvNext | โ | โ | โ
| โ
| โ |
| CTRL | โ
| โ | โ
| โ
| โ |
| CvT | โ | โ | โ
| โ | โ |
| Data2VecAudio | โ | โ | โ
| โ | โ |
| Data2VecText | โ | โ | โ
| โ | โ |
| Data2VecVision | โ | โ | โ
| โ
| โ |
| DeBERTa | โ
| โ
| โ
| โ
| โ |
| DeBERTa-v2 | โ
| โ
| โ
| โ
| โ |
| Decision Transformer | โ | โ | โ
| โ | โ |
| DeiT | โ | โ | โ
| โ | โ |
| DETR | โ | โ | โ
| โ | โ |
| DistilBERT | โ
| โ
| โ
| โ
| โ
|
| DPR | โ
| โ
| โ
| โ
| โ |
| DPT | โ | โ | โ
| โ | โ |
| ELECTRA | โ
| โ
| โ
| โ
| โ
|
| Encoder decoder | โ | โ | โ
| โ
| โ
|
| FairSeq Machine-Translation | โ
| โ | โ
| โ | โ |
| FlauBERT | โ
| โ | โ
| โ
| โ |
| Flava | โ | โ | โ
| โ | โ |
| FNet | โ
| โ
| โ
| โ | โ |
| Funnel Transformer | โ
| โ
| โ
| โ
| โ |
| GLPN | โ | โ | โ
| โ | โ |
| GPT Neo | โ | โ | โ
| โ | โ
|
| GPT NeoX | โ | โ
| โ
| โ | โ |
| GPT-J | โ | โ | โ
| โ
| โ
|
| Hubert | โ | โ | โ
| โ
| โ |
| I-BERT | โ | โ | โ
| โ | โ |
| ImageGPT | โ | โ | โ
| โ | โ |
| LayoutLM | โ
| โ
| โ
| โ
| โ |
| LayoutLMv2 | โ
| โ
| โ
| โ | โ |
| LayoutLMv3 | โ
| โ
| โ
| โ
| โ |
| LED | โ
| โ
| โ
| โ
| โ |
| Longformer | โ
| โ
| โ
| โ
| โ |
| LUKE | โ
| โ | โ
| โ | โ |
| LXMERT | โ
| โ
| โ
| โ
| โ |
| M2M100 | โ
| โ | โ
| โ | โ |
| Marian | โ
| โ | โ
| โ
| โ
|
| MaskFormer | โ | โ | โ
| โ | โ |
| mBART | โ
| โ
| โ
| โ
| โ
|
| MegatronBert | โ | โ | โ
| โ | โ |
| MobileBERT | โ
| โ
| โ
| โ
| โ |
| MPNet | โ
| โ
| โ
| โ
| โ |
| mT5 | โ
| โ
| โ
| โ
| โ
|
| Nystromformer | โ | โ | โ
| โ | โ |
| OpenAI GPT | โ
| โ
| โ
| โ
| โ |
| OpenAI GPT-2 | โ
| โ
| โ
| โ
| โ
|
| OPT | โ | โ | โ
| โ | โ |
| Pegasus | โ
| โ
| โ
| โ
| โ
|
| Perceiver | โ
| โ | โ
| โ | โ |
| PLBart | โ
| โ | โ
| โ | โ |
| PoolFormer | โ | โ | โ
| โ | โ |
| ProphetNet | โ
| โ | โ
| โ | โ |
| QDQBert | โ | โ | โ
| โ | โ |
| RAG | โ
| โ | โ
| โ
| โ |
| Realm | โ
| โ
| โ
| โ | โ |
| Reformer | โ
| โ
| โ
| โ | โ |
| RegNet | โ | โ | โ
| โ
| โ
|
| RemBERT | โ
| โ
| โ
| โ
| โ |
| ResNet | โ | โ | โ
| โ
| โ
|
| RetriBERT | โ
| โ
| โ
| โ | โ |
| RoBERTa | โ
| โ
| โ
| โ
| โ
|
| RoFormer | โ
| โ
| โ
| โ
| โ
|
| SegFormer | โ | โ | โ
| โ | โ |
| SEW | โ | โ | โ
| โ | โ |
| SEW-D | โ | โ | โ
| โ | โ |
| Speech Encoder decoder | โ | โ | โ
| โ | โ
|
| Speech2Text | โ
| โ | โ
| โ
| โ |
| Speech2Text2 | โ
| โ | โ | โ | โ |
| Splinter | โ
| โ
| โ
| โ | โ |
| SqueezeBERT | โ
| โ
| โ
| โ | โ |
| Swin | โ | โ | โ
| โ
| โ |
| T5 | โ
| โ
| โ
| โ
| โ
|
| TAPAS | โ
| โ | โ
| โ
| โ |
| Trajectory Transformer | โ | โ | โ
| โ | โ |
| Transformer-XL | โ
| โ | โ
| โ
| โ |
| TrOCR | โ | โ | โ
| โ | โ |
| UniSpeech | โ | โ | โ
| โ | โ |
| UniSpeechSat | โ | โ | โ
| โ | โ |
| VAN | โ | โ | โ
| โ | โ |
| ViLT | โ | โ | โ
| โ | โ |
| Vision Encoder decoder | โ | โ | โ
| โ
| โ
|
| VisionTextDualEncoder | โ | โ | โ
| โ | โ
|
| VisualBert | โ | โ | โ
| โ | โ |
| ViT | โ | โ | โ
| โ
| โ
|
| ViTMAE | โ | โ | โ
| โ
| โ |
| Wav2Vec2 | โ
| โ | โ
| โ
| โ
|
| Wav2Vec2-Conformer | โ | โ | โ
| โ | โ |
| WavLM | โ | โ | โ
| โ | โ |
| XGLM | โ
| โ
| โ
| โ | โ
|
| XLM | โ
| โ | โ
| โ
| โ |
| XLM-RoBERTa | โ
| โ
| โ
| โ
| โ
|
| XLM-RoBERTa-XL | โ | โ | โ
| โ | โ |
| XLMProphetNet | โ
| โ | โ
| โ | โ |
| XLNet | โ
| โ
| โ
| โ
| โ |
| YOLOS | โ | โ | โ
| โ | โ |
| YOSO | โ | โ | โ
| โ | โ |
<!-- End table-->
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/custom_models.md
|
<!--Copyright 2020 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Condividere modelli personalizzati
La libreria ๐ค Transformers รจ studiata per essere facilmente estendibile. Il codice di ogni modello รจ interamente
situato in una sottocartella del repository senza alcuna astrazione, perciรฒ puoi facilmente copiare il file di un
modello e modificarlo in base ai tuoi bisogni.
Se stai scrivendo un nuovo modello, potrebbe essere piรน semplice iniziare da zero. In questo tutorial, ti mostreremo
come scrivere un modello personalizzato e la sua configurazione in modo che possa essere utilizzato allโinterno di
Transformers, e come condividerlo con la community (assieme al relativo codice) cosรฌ che tutte le persone possano usarlo, anche
se non presente nella libreria ๐ค Transformers.
Illustriamo tutto questo su un modello ResNet, avvolgendo la classe ResNet della
[libreria timm](https://github.com/rwightman/pytorch-image-models) in un [`PreTrainedModel`].
## Scrivere una configurazione personalizzata
Prima di iniziare a lavorare al modello, scriviamone la configurazione. La configurazione di un modello รจ un oggetto
che contiene tutte le informazioni necessarie per la build del modello. Come vedremo nella prossima sezione, il
modello puรฒ soltanto essere inizializzato tramite `config`, per cui dovremo rendere tale oggetto piรน completo possibile.
Nel nostro esempio, prenderemo un paio di argomenti della classe ResNet che potremmo voler modificare.
Configurazioni differenti ci daranno quindi i differenti possibili tipi di ResNet. Salveremo poi questi argomenti,
dopo averne controllato la validitร .
```python
from transformers import PretrainedConfig
from typing import List
class ResnetConfig(PretrainedConfig):
model_type = "resnet"
def __init__(
self,
block_type="bottleneck",
layers: List[int] = [3, 4, 6, 3],
num_classes: int = 1000,
input_channels: int = 3,
cardinality: int = 1,
base_width: int = 64,
stem_width: int = 64,
stem_type: str = "",
avg_down: bool = False,
**kwargs,
):
if block_type not in ["basic", "bottleneck"]:
raise ValueError(f"`block_type` must be 'basic' or bottleneck', got {block_type}.")
if stem_type not in ["", "deep", "deep-tiered"]:
raise ValueError(f"`stem_type` must be '', 'deep' or 'deep-tiered', got {stem_type}.")
self.block_type = block_type
self.layers = layers
self.num_classes = num_classes
self.input_channels = input_channels
self.cardinality = cardinality
self.base_width = base_width
self.stem_width = stem_width
self.stem_type = stem_type
self.avg_down = avg_down
super().__init__(**kwargs)
```
Le tre cose piรน importanti da ricordare quando scrivi le tue configurazioni sono le seguenti:
- Devi ereditare da `Pretrainedconfig`,
- Il metodo `__init__` del tuo `Pretrainedconfig` deve accettare i kwargs,
- I `kwargs` devono essere passati alla superclass `__init__`
Lโereditร รจ importante per assicurarsi di ottenere tutte le funzionalitร della libreria ๐ค transformers,
mentre gli altri due vincoli derivano dal fatto che un `Pretrainedconfig` ha piรน campi di quelli che stai settando.
Quando ricarichi una config da un metodo `from_pretrained`, questi campi devono essere accettati dalla tua config e
poi inviati alla superclasse.
Definire un `model_type` per la tua configurazione (qua `model_type = โresnetโ`) non รจ obbligatorio, a meno che tu
non voglia registrare il modello con le classi Auto (vedi l'ultima sezione).
Una volta completato, puoi facilmente creare e salvare la tua configurazione come faresti con ogni altra configurazione
di modelli della libreria. Ecco come possiamo creare la config di un resnet50d e salvarlo:
```py
resnet50d_config = ResnetConfig(block_type="bottleneck", stem_width=32, stem_type="deep", avg_down=True)
resnet50d_config.save_pretrained("custom-resnet")
```
Questo salverร un file chiamato `config.json` all'interno della cartella `custom-resnet`. Potrai poi ricaricare la tua
config con il metodo `from_pretrained`.
```py
resnet50d_config = ResnetConfig.from_pretrained("custom-resnet")
```
Puoi anche usare qualunque altro metodo della classe [`PretrainedConfig`], come [`~PretrainedConfig.push_to_hub`]
per caricare direttamente la tua configurazione nell'hub.
## Scrivere un modello personalizzato
Ora che abbiamo la nostra configurazione ResNet, possiamo continuare a scrivere il modello. In realtร , ne scriveremo
due: uno che estrae le features nascoste da una batch di immagini (come [`BertModel`]) e uno che รจ utilizzabile per
la classificazione di immagini (come [`BertModelForSequenceClassification`]).
Come abbiamo menzionato in precedenza, scriveremo soltanto un wrapper del modello, per mantenerlo semplice ai fini di
questo esempio. L'unica cosa che dobbiamo fare prima di scrivere questa classe รจ una mappatura fra i tipi di blocco e
le vere classi dei blocchi. Successivamente il modello รจ definito tramite la configurazione, passando tutto quanto alla
classe `ResNet`.
```py
from transformers import PreTrainedModel
from timm.models.resnet import BasicBlock, Bottleneck, ResNet
from .configuration_resnet import ResnetConfig
BLOCK_MAPPING = {"basic": BasicBlock, "bottleneck": Bottleneck}
class ResnetModel(PreTrainedModel):
config_class = ResnetConfig
def __init__(self, config):
super().__init__(config)
block_layer = BLOCK_MAPPING[config.block_type]
self.model = ResNet(
block_layer,
config.layers,
num_classes=config.num_classes,
in_chans=config.input_channels,
cardinality=config.cardinality,
base_width=config.base_width,
stem_width=config.stem_width,
stem_type=config.stem_type,
avg_down=config.avg_down,
)
def forward(self, tensor):
return self.model.forward_features(tensor)
```
Per il modello che classificherร le immagini, cambiamo soltanto il metodo forward:
```py
import torch
class ResnetModelForImageClassification(PreTrainedModel):
config_class = ResnetConfig
def __init__(self, config):
super().__init__(config)
block_layer = BLOCK_MAPPING[config.block_type]
self.model = ResNet(
block_layer,
config.layers,
num_classes=config.num_classes,
in_chans=config.input_channels,
cardinality=config.cardinality,
base_width=config.base_width,
stem_width=config.stem_width,
stem_type=config.stem_type,
avg_down=config.avg_down,
)
def forward(self, tensor, labels=None):
logits = self.model(tensor)
if labels is not None:
loss = torch.nn.cross_entropy(logits, labels)
return {"loss": loss, "logits": logits}
return {"logits": logits}
```
Nota come, in entrambi i casi, ereditiamo da `PreTrainedModel` e chiamiamo l'inizializzazione della superclasse
con il metodo `config` (un po' come quando scrivi un normale `torch.nn.Module`). La riga che imposta la `config_class`
non รจ obbligatoria, a meno che tu non voglia registrare il modello con le classi Auto (vedi l'ultima sezione).
<Tip>
Se il tuo modello รจ molto simile a un modello all'interno della libreria, puoi ri-usare la stessa configurazione di quel modello.
</Tip>
Puoi fare in modo che il tuo modello restituisca in output qualunque cosa tu voglia, ma far restituire un dizionario
come abbiamo fatto per `ResnetModelForImageClassification`, con la funzione di perdita inclusa quando vengono passate le labels,
renderร il tuo modello direttamente utilizzabile all'interno della classe [`Trainer`]. Utilizzare altri formati di output va bene
se hai in progetto di utilizzare un tuo loop di allenamento, o se utilizzerai un'altra libreria per l'addestramento.
Ora che abbiamo la classe del nostro modello, creiamone uno:
```py
resnet50d = ResnetModelForImageClassification(resnet50d_config)
```
Ribadiamo, puoi usare qualunque metodo dei [`PreTrainedModel`], come [`~PreTrainedModel.save_pretrained`] o
[`~PreTrainedModel.push_to_hub`]. Utilizzeremo quest'ultimo nella prossima sezione, e vedremo come caricare i pesi del
modello assieme al codice del modello stesso. Ma prima, carichiamo alcuni pesi pre-allenati all'interno del nostro modello.
Nel tuo caso specifico, probabilmente allenerai il tuo modello sui tuoi dati. Per velocizzare in questo tutorial,
utilizzeremo la versione pre-allenata del resnet50d. Dato che il nostro modello รจ soltanto un wrapper attorno a quel modello,
sarร facile trasferirne i pesi:
```py
import timm
pretrained_model = timm.create_model("resnet50d", pretrained=True)
resnet50d.model.load_state_dict(pretrained_model.state_dict())
```
Vediamo adesso come assicurarci che quando facciamo [`~PreTrainedModel.save_pretrained`] o [`~PreTrainedModel.push_to_hub`],
il codice del modello venga salvato.
## Inviare il codice all'Hub
<Tip warning={true}>
Questa API รจ sperimentale e potrebbe avere alcuni cambiamenti nei prossimi rilasci.
</Tip>
Innanzitutto, assicurati che il tuo modello sia completamente definito in un file `.py`. Puรฒ sfruttare import relativi
ad altri file, purchรจ questi siano nella stessa directory (non supportiamo ancora sotto-moduli per questa funzionalitร ).
Per questo esempio, definiremo un file `modeling_resnet.py` e un file `configuration_resnet.py` in una cartella dell'attuale
working directory chiamata `resnet_model`. Il file configuration contiene il codice per `ResnetConfig` e il file modeling
contiene il codice di `ResnetModel` e `ResnetModelForImageClassification`.
```
.
โโโ resnet_model
โโโ __init__.py
โโโ configuration_resnet.py
โโโ modeling_resnet.py
```
Il file `__init__.py` puรฒ essere vuoto, serve solo perchรจ Python capisca che `resnet_model` puรฒ essere utilizzato come un modulo.
<Tip warning={true}>
Se stai copiando i file relativi alla modellazione della libreria, dovrai sostituire tutti gli import relativi in cima al file con import del
pacchetto `transformers`.
</Tip>
Nota che puoi ri-utilizzare (o usare come sottoclassi) un modello/configurazione esistente.
Per condividere il tuo modello con la community, segui questi passi: prima importa il modello ResNet e la sua configurazione
dai nuovi file creati:
```py
from resnet_model.configuration_resnet import ResnetConfig
from resnet_model.modeling_resnet import ResnetModel, ResnetModelForImageClassification
```
Dopodichรจ dovrai dire alla libreria che vuoi copiare i file con il codice di quegli oggetti quando utilizzi il metodo
`save_pretrained` e registrarli in modo corretto con una Auto classe (specialmente per i modelli). Utilizza semplicemente:
```py
ResnetConfig.register_for_auto_class()
ResnetModel.register_for_auto_class("AutoModel")
ResnetModelForImageClassification.register_for_auto_class("AutoModelForImageClassification")
```
Nota che non c'รจ bisogno di specificare una Auto classe per la configurazione (c'รจ solo una Auto classe per le configurazioni,
[`AutoConfig`], ma รจ diversa per i modelli). Il tuo modello personalizato potrebbe essere utilizzato per diverse tasks,
per cui devi specificare quale delle classi Auto รจ quella corretta per il tuo modello.
Successivamente, creiamo i modelli e la config come abbiamo fatto in precedenza:
```py
resnet50d_config = ResnetConfig(block_type="bottleneck", stem_width=32, stem_type="deep", avg_down=True)
resnet50d = ResnetModelForImageClassification(resnet50d_config)
pretrained_model = timm.create_model("resnet50d", pretrained=True)
resnet50d.model.load_state_dict(pretrained_model.state_dict())
```
Adesso, per inviare il modello all'Hub, assicurati di aver effettuato l'accesso. Lancia dal tuo terminale:
```bash
huggingface-cli login
```
O da un notebook:
```py
from huggingface_hub import notebook_login
notebook_login()
```
Potrai poi inviare il tutto sul tuo profilo (o di un'organizzazione di cui fai parte) in questo modo:
```py
resnet50d.push_to_hub("custom-resnet50d")
```
Oltre ai pesi del modello e alla configurazione in formato json, questo ha anche copiato i file `.py` modeling e
configuration all'interno della cartella `custom-resnet50d` e ha caricato i risultati sull'Hub. Puoi controllare
i risultati in questa [model repo](https://huggingface.co/sgugger/custom-resnet50d).
Puoi controllare il tutorial di condivisione [tutorial di condivisione](model_sharing) per piรน informazioni sul
metodo con cui inviare all'Hub.
## Usare un modello con codice personalizzato
Puoi usare ogni configurazione, modello o tokenizer con file di codice personalizzati nella sua repository
con le classi Auto e il metodo `from_pretrained`. Tutti i files e il codice caricati sull'Hub sono scansionati da malware
(fai riferimento alla documentazione [Hub security](https://huggingface.co/docs/hub/security#malware-scanning) per piรน informazioni),
ma dovresti comunque assicurarti dell'affidabilitร del codice e dell'autore per evitare di eseguire codice dannoso sulla tua macchina.
Imposta `trust_remote_code=True` per usare un modello con codice personalizzato:
```py
from transformers import AutoModelForImageClassification
model = AutoModelForImageClassification.from_pretrained("sgugger/custom-resnet50d", trust_remote_code=True)
```
Inoltre, raccomandiamo fortemente di passare un hash del commit come `revision` per assicurarti che le autrici o gli autori del modello
non abbiano modificato il codice con alcune nuove righe dannose (a meno che non ti fidi completamente della fonte):
```py
commit_hash = "ed94a7c6247d8aedce4647f00f20de6875b5b292"
model = AutoModelForImageClassification.from_pretrained(
"sgugger/custom-resnet50d", trust_remote_code=True, revision=commit_hash
)
```
Nota che quando cerchi la storia dei commit della repo del modello sull'Hub, c'รจ un bottone con cui facilmente copiare il
commit hash di ciascun commit.
## Registrare un modello con codice personalizzato nelle classi Auto
Se stai scrivendo una libreria che estende ๐ค Transformers, potresti voler estendere le classi Auto per includere il tuo modello.
Questo รจ diverso dall'inviare codice nell'Hub: gli utenti dovranno importare la tua libreria per ottenere il modello personalizzato
(anzichรจ scaricare automaticamente il modello dall'Hub).
Finchรจ il tuo file di configurazione ha un attributo `model_type` diverso dai model types esistenti, e finchรจ le tue
classi modello hanno i corretti attributi `config_class`, potrai semplicemente aggiungerli alle classi Auto come segue:
```py
from transformers import AutoConfig, AutoModel, AutoModelForImageClassification
AutoConfig.register("resnet", ResnetConfig)
AutoModel.register(ResnetConfig, ResnetModel)
AutoModelForImageClassification.register(ResnetConfig, ResnetModelForImageClassification)
```
Nota che il primo argomento utilizzato quando registri la configurazione di un modello personalizzato con [`AutoConfig`]
deve corrispondere al `model_type` della tua configurazione personalizzata, ed il primo argomento utilizzato quando
registri i tuoi modelli personalizzati in una qualunque classe Auto del modello deve corrispondere alla `config_class`
di quei modelli.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/perf_hardware.md
|
<!---
Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Hardware ottimizzato per l'addestramento
L'hardware utilizzato per eseguire l'addestramento del modello e l'inferenza puรฒ avere un grande effetto sulle prestazioni. Per un analisi approfondita delle GPUs, assicurati di dare un'occhiata all'eccellente [blog post](https://timdettmers.com/2020/09/07/which-gpu-for-deep-learning/) di Tim Dettmer.
Diamo un'occhiata ad alcuni consigli pratici per la configurazione della GPU.
## GPU
Quando si addestrano modelli piรน grandi ci sono essenzialmente tre opzioni:
- GPUs piu' grandi
- Piu' GPUs
- Piu' CPU e piu' NVMe (scaricato da [DeepSpeed-Infinity](main_classes/deepspeed#nvme-support))
Iniziamo dal caso in cui ci sia una singola GPU.
### Potenza e Raffreddamento
Se hai acquistato una costosa GPU di fascia alta, assicurati di darle la potenza corretta e un raffreddamento sufficiente.
**Potenza**:
Alcune schede GPU consumer di fascia alta hanno 2 e talvolta 3 prese di alimentazione PCI-E a 8 pin. Assicurati di avere tanti cavi PCI-E a 8 pin indipendenti da 12 V collegati alla scheda quante sono le prese. Non utilizzare le 2 fessure a un'estremitร dello stesso cavo (noto anche come cavo a spirale). Cioรจ se hai 2 prese sulla GPU, vuoi 2 cavi PCI-E a 8 pin che vanno dall'alimentatore alla scheda e non uno che abbia 2 connettori PCI-E a 8 pin alla fine! In caso contrario, non otterrai tutte le prestazioni ufficiali.
Ciascun cavo di alimentazione PCI-E a 8 pin deve essere collegato a una guida da 12 V sul lato dell'alimentatore e puรฒ fornire fino a 150 W di potenza.
Alcune altre schede possono utilizzare connettori PCI-E a 12 pin e questi possono fornire fino a 500-600 W di potenza.
Le schede di fascia bassa possono utilizzare connettori a 6 pin, che forniscono fino a 75 W di potenza.
Inoltre vuoi un alimentatore (PSU) di fascia alta che abbia una tensione stabile. Alcuni PSU di qualitร inferiore potrebbero non fornire alla scheda la tensione stabile di cui ha bisogno per funzionare al massimo.
E ovviamente l'alimentatore deve avere abbastanza Watt inutilizzati per alimentare la scheda.
**Raffreddamento**:
Quando una GPU si surriscalda, inizierร a rallentare e non fornirร le prestazioni mssimali e potrebbe persino spegnersi se diventasse troppo calda.
ร difficile dire l'esatta temperatura migliore a cui aspirare quando una GPU รจ molto caricata, ma probabilmente qualsiasi cosa al di sotto di +80ยฐC va bene, ma piรน bassa รจ meglio - forse 70-75ยฐC รจ un intervallo eccellente in cui trovarsi. ร probabile che il rallentamento inizi a circa 84-90ยฐC. Ma oltre alla limitazione delle prestazioni, una temperatura molto elevata prolungata รจ probabile che riduca la durata di una GPU.
Diamo quindi un'occhiata a uno degli aspetti piรน importanti quando si hanno piรน GPU: la connettivitร .
### Connettivitร multi-GPU
Se utilizzi piรน GPU, il modo in cui le schede sono interconnesse puรฒ avere un enorme impatto sul tempo totale di allenamento. Se le GPU si trovano sullo stesso nodo fisico, puoi eseguire:
```
nvidia-smi topo -m
```
e ti dirร come sono interconnesse le GPU. Su una macchina con doppia GPU e collegata a NVLink, molto probabilmente vedrai qualcosa del tipo:
```
GPU0 GPU1 CPU Affinity NUMA Affinity
GPU0 X NV2 0-23 N/A
GPU1 NV2 X 0-23 N/A
```
su una macchina diversa senza NVLink potremmo vedere:
```
GPU0 GPU1 CPU Affinity NUMA Affinity
GPU0 X PHB 0-11 N/A
GPU1 PHB X 0-11 N/A
```
Il rapporto include questa legenda:
```
X = Self
SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
PIX = Connection traversing at most a single PCIe bridge
NV# = Connection traversing a bonded set of # NVLinks
```
Quindi il primo rapporto `NV2` ci dice che le GPU sono interconnesse con 2 NVLinks e nel secondo report `PHB` abbiamo una tipica configurazione PCIe+Bridge a livello di consumatore.
Controlla che tipo di connettivitร hai sulla tua configurazione. Alcuni di questi renderanno la comunicazione tra le carte piรน veloce (es. NVLink), altri piรน lenta (es. PHB).
A seconda del tipo di soluzione di scalabilitร utilizzata, la velocitร di connettivitร potrebbe avere un impatto maggiore o minore. Se le GPU devono sincronizzarsi raramente, come in DDP, l'impatto di una connessione piรน lenta sarร meno significativo. Se le GPU devono scambiarsi messaggi spesso, come in ZeRO-DP, una connettivitร piรน veloce diventa estremamente importante per ottenere un addestramento piรน veloce.
#### NVlink
[NVLink](https://en.wikipedia.org/wiki/NVLink) รจ un collegamento di comunicazione a corto raggio multilinea seriale basato su cavo sviluppato da Nvidia.
Ogni nuova generazione fornisce una larghezza di banda piรน veloce, ad es. ecco una citazione da [Nvidia Ampere GA102 GPU Architecture](https://www.nvidia.com/content/dam/en-zz/Solutions/geforce/ampere/pdf/NVIDIA-ampere-GA102-GPU-Architecture-Whitepaper-V1.pdf):
> Third-Generation NVLinkยฎ
> GA102 GPUs utilize NVIDIAโs third-generation NVLink interface, which includes four x4 links,
> with each link providing 14.0625 GB/sec bandwidth in each direction between two GPUs. Four
> links provide 56.25 GB/sec bandwidth in each direction, and 112.5 GB/sec total bandwidth
> between two GPUs. Two RTX 3090 GPUs can be connected together for SLI using NVLink.
> (Note that 3-Way and 4-Way SLI configurations are not supported.)
Quindi piรน `X` si ottiene nel rapporto di `NVX` nell'output di `nvidia-smi topo -m`, meglio รจ. La generazione dipenderร dall'architettura della tua GPU.
Confrontiamo l'esecuzione di un training del modello di linguaggio gpt2 su un piccolo campione di wikitext
I risultati sono:
| NVlink | Time |
| ----- | ---: |
| Y | 101s |
| N | 131s |
Puoi vedere che NVLink completa l'addestramento circa il 23% piรน velocemente. Nel secondo benchmark utilizziamo `NCCL_P2P_DISABLE=1` per dire alle GPU di non utilizzare NVLink.
Ecco il codice benchmark completo e gli output:
```bash
# DDP w/ NVLink
rm -r /tmp/test-clm; CUDA_VISIBLE_DEVICES=0,1 torchrun \
--nproc_per_node 2 examples/pytorch/language-modeling/run_clm.py --model_name_or_path gpt2 \
--dataset_name wikitext --dataset_config_name wikitext-2-raw-v1 --do_train \
--output_dir /tmp/test-clm --per_device_train_batch_size 4 --max_steps 200
{'train_runtime': 101.9003, 'train_samples_per_second': 1.963, 'epoch': 0.69}
# DDP w/o NVLink
rm -r /tmp/test-clm; CUDA_VISIBLE_DEVICES=0,1 NCCL_P2P_DISABLE=1 torchrun \
--nproc_per_node 2 examples/pytorch/language-modeling/run_clm.py --model_name_or_path gpt2 \
--dataset_name wikitext --dataset_config_name wikitext-2-raw-v1 --do_train
--output_dir /tmp/test-clm --per_device_train_batch_size 4 --max_steps 200
{'train_runtime': 131.4367, 'train_samples_per_second': 1.522, 'epoch': 0.69}
```
Hardware: 2x TITAN RTX 24GB each + NVlink with 2 NVLinks (`NV2` in `nvidia-smi topo -m`)
Software: `pytorch-1.8-to-be` + `cuda-11.0` / `transformers==4.3.0.dev0`
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/perf_train_tpu.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Addestramento su TPU
<Tip>
Nota: Molte delle strategie introdotte nella [sezione sulla GPU singola](perf_train_gpu_one) (come mixed precision training o gradient accumulation) e [sezione multi-GPU](perf_train_gpu_many) sono generiche e applicabili all'addestramento di modelli in generale quindi assicurati di dargli un'occhiata prima di immergerti in questa sezione.
</Tip>
Questo documento sarร presto completato con informazioni su come effettuare la formazione su TPU.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/quicktour.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Quick tour
[[open-in-colab]]
Entra in azione con ๐ค Transformers! Inizia utilizzando [`pipeline`] per un'inferenza veloce, carica un modello pre-allenato e un tokenizer con una [AutoClass](./model_doc/auto) per risolvere i tuoi compiti legati a testo, immagini o audio.
<Tip>
Tutti gli esempi di codice presenti in questa documentazione hanno un pulsante in alto a sinistra che permette di selezionare tra PyTorch e TensorFlow. Se
questo non รจ presente, ci si aspetta che il codice funzioni per entrambi i backend senza alcun cambiamento.
</Tip>
## Pipeline
[`pipeline`] รจ il modo piรน semplice per utilizzare un modello pre-allenato per un dato compito.
<Youtube id="tiZFewofSLM"/>
La [`pipeline`] supporta molti compiti comuni:
**Testo**:
* Analisi del Sentimento (Sentiment Analysis, in inglese): classifica la polaritร di un testo dato.
* Generazione del Testo (Text Generation, in inglese): genera del testo a partire da un dato input.
* Riconoscimento di Entitร (Name Entity Recognition o NER, in inglese): etichetta ogni parola con l'entitร che questa rappresenta (persona, data, luogo, ecc.).
* Rispondere a Domande (Question answering, in inglese): estrae la risposta da un contesto, dato del contesto e una domanda.
* Riempimento di Maschere (Fill-mask, in inglese): riempie gli spazi mancanti in un testo che ha parole mascherate.
* Riassumere (Summarization, in inglese): genera una sintesi di una lunga sequenza di testo o di un documento.
* Traduzione (Translation, in inglese): traduce un testo in un'altra lingua.
* Estrazione di Caratteristiche (Feature Extraction, in inglese): crea un tensore che rappresenta un testo.
**Immagini**:
* Classificazione di Immagini (Image Classification, in inglese): classifica un'immagine.
* Segmentazione di Immagini (Image Segmentation, in inglese): classifica ogni pixel di un'immagine.
* Rilevazione di Oggetti (Object Detection, in inglese): rileva oggetti all'interno di un'immagine.
**Audio**:
* Classificazione di Audio (Audio Classification, in inglese): assegna un'etichetta ad un segmento di audio dato.
* Riconoscimento Vocale Automatico (Automatic Speech Recognition o ASR, in inglese): trascrive il contenuto di un audio dato in un testo.
<Tip>
Per maggiori dettagli legati alla [`pipeline`] e ai compiti ad essa associati, fai riferimento alla documentazione [qui](./main_classes/pipelines).
</Tip>
### Utilizzo della Pipeline
Nel seguente esempio, utilizzerai la [`pipeline`] per l'analisi del sentimento.
Installa le seguenti dipendenze se non lo hai giร fatto:
<frameworkcontent>
<pt>
```bash
pip install torch
```
</pt>
<tf>
```bash
pip install tensorflow
```
</tf>
</frameworkcontent>
Importa [`pipeline`] e specifica il compito che vuoi completare:
```py
>>> from transformers import pipeline
>>> classificatore = pipeline("sentiment-analysis", model="MilaNLProc/feel-it-italian-sentiment")
```
La pipeline scarica e salva il [modello pre-allenato](https://huggingface.co/MilaNLProc/feel-it-italian-sentiment) e il tokenizer per l'analisi del sentimento. Se non avessimo scelto un modello, la pipeline ne avrebbe scelto uno di default. Ora puoi utilizzare il `classifier` sul tuo testo obiettivo:
```py
>>> classificatore("Siamo molto felici di mostrarti la libreria ๐ค Transformers.")
[{'label': 'positive', 'score': 0.9997}]
```
Per piรน di una frase, passa una lista di frasi alla [`pipeline`] la quale restituirร una lista di dizionari:
```py
>>> risultati = classificatore(
... ["Siamo molto felici di mostrarti la libreria ๐ค Transformers.", "Speriamo te non la odierai."]
... )
>>> for risultato in risultati:
... print(f"etichetta: {risultato['label']}, con punteggio: {round(risultato['score'], 4)}")
etichetta: positive, con punteggio: 0.9998
etichetta: negative, con punteggio: 0.9998
```
La [`pipeline`] puรฒ anche iterare su un dataset intero. Inizia installando la libreria [๐ค Datasets](https://huggingface.co/docs/datasets/):
```bash
pip install datasets
```
Crea una [`pipeline`] con il compito che vuoi risolvere e con il modello che vuoi utilizzare.
```py
>>> import torch
>>> from transformers import pipeline
>>> riconoscitore_vocale = pipeline(
... "automatic-speech-recognition", model="radiogroup-crits/wav2vec2-xls-r-1b-italian-doc4lm-5gram"
... )
```
Poi, carica un dataset (vedi ๐ค Datasets [Quick Start](https://huggingface.co/docs/datasets/quickstart) per maggiori dettagli) sul quale vuoi iterare. Per esempio, carichiamo il dataset [MInDS-14](https://huggingface.co/datasets/PolyAI/minds14):
```py
>>> from datasets import load_dataset, Audio
>>> dataset = load_dataset("PolyAI/minds14", name="it-IT", split="train") # doctest: +IGNORE_RESULT
```
Dobbiamo assicurarci che la frequenza di campionamento del set di dati corrisponda alla frequenza di campionamento con cui รจ stato addestrato `radiogroup-crits/wav2vec2-xls-r-1b-italian-doc4lm-5gram`.
```py
>>> dataset = dataset.cast_column("audio", Audio(sampling_rate=riconoscitore_vocale.feature_extractor.sampling_rate))
```
I file audio vengono caricati automaticamente e ri-campionati quando chiamiamo la colonna "audio".
Estraiamo i vettori delle forme d'onda grezze delle prime 4 osservazioni e passiamoli come lista alla pipeline:
```py
>>> risultato = riconoscitore_vocale(dataset[:4]["audio"])
>>> print([d["text"] for d in risultato])
['dovrei caricare dei soldi sul mio conto corrente', 'buongiorno e senza vorrei depositare denaro sul mio conto corrente come devo fare per cortesia', 'sรฌ salve vorrei depositare del denaro sul mio conto', 'e buon pomeriggio vorrei depositare dei soldi sul mio conto bancario volleo sapere come posso fare se e posso farlo online ed un altro conto o andandoo tramite bancomut']
```
Per un dataset piรน grande dove gli input sono di dimensione maggiore (come nel parlato/audio o nella visione), dovrai passare un generatore al posto di una lista che carica tutti gli input in memoria. Guarda la [documentazione della pipeline](./main_classes/pipelines) per maggiori informazioni.
### Utilizzare un altro modello e tokenizer nella pipeline
La [`pipeline`] puรฒ ospitare qualsiasi modello del [Model Hub](https://huggingface.co/models), rendendo semplice l'adattamento della [`pipeline`] per altri casi d'uso. Per esempio, se si vuole un modello capace di trattare testo in francese, usa i tag presenti nel Model Hub in modo da filtrare per ottenere un modello appropriato. Il miglior risultato filtrato restituisce un modello multi-lingua [BERT model](https://huggingface.co/nlptown/bert-base-multilingual-uncased-sentiment) fine-tuned per l'analisi del sentimento. Ottimo, utilizziamo questo modello!
```py
>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
```
<frameworkcontent>
<pt>
Usa [`AutoModelForSequenceClassification`] e [`AutoTokenizer`] per caricare il modello pre-allenato e il suo tokenizer associato (maggiori informazioni su una `AutoClass` in seguito):
```py
>>> from transformers import AutoTokenizer, AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained(model_name)
>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
```
</pt>
<tf>
Usa [`TFAutoModelForSequenceClassification`] e [`AutoTokenizer`] per caricare il modello pre-allenato e il suo tokenizer associato (maggiori informazioni su una `TFAutoClass` in seguito):
```py
>>> from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
>>> model = TFAutoModelForSequenceClassification.from_pretrained(model_name)
>>> tokenizer = AutoTokenizer.from_pretrained(model_name)
```
</tf>
</frameworkcontent>
Poi puoi specificare il modello e il tokenizer nella [`pipeline`], e applicare il `classifier` sul tuo testo obiettivo:
```py
>>> classifier = pipeline("sentiment-analysis", model=model, tokenizer=tokenizer)
>>> classifier("Nous sommes trรจs heureux de vous prรฉsenter la bibliothรจque ๐ค Transformers.")
[{'label': '5 stars', 'score': 0.7273}]
```
Se non riesci a trovare un modello per il tuo caso d'uso, dovrai fare fine-tuning di un modello pre-allenato sui tuoi dati. Dai un'occhiata al nostro tutorial [fine-tuning tutorial](./training) per imparare come. Infine, dopo che hai completato il fine-tuning del tuo modello pre-allenato, considera per favore di condividerlo (vedi il tutorial [qui](./model_sharing)) con la comunitร sul Model Hub per democratizzare l'NLP! ๐ค
## AutoClass
<Youtube id="AhChOFRegn4"/>
Al suo interno, le classi [`AutoModelForSequenceClassification`] e [`AutoTokenizer`] lavorano assieme per dare potere alla [`pipeline`]. Una [AutoClass](./model_doc/auto) รจ una scorciatoia che automaticamente recupera l'architettura di un modello pre-allenato a partire dal suo nome o path. Hai solo bisogno di selezionare la `AutoClass` appropriata per il tuo compito e il suo tokenizer associato con [`AutoTokenizer`].
Ritorniamo al nostro esempio e vediamo come puoi utilizzare la `AutoClass` per replicare i risultati della [`pipeline`].
### AutoTokenizer
Un tokenizer รจ responsabile dell'elaborazione del testo in modo da trasformarlo in un formato comprensibile dal modello. Per prima cosa, il tokenizer dividerร il testo in parole chiamate *token*. Ci sono diverse regole che governano il processo di tokenizzazione, tra cui come dividere una parola e a quale livello (impara di piรน sulla tokenizzazione [qui](./tokenizer_summary)). La cosa piรน importante da ricordare comunque รจ che hai bisogno di inizializzare il tokenizer con lo stesso nome del modello in modo da assicurarti che stai utilizzando le stesse regole di tokenizzazione con cui il modello รจ stato pre-allenato.
Carica un tokenizer con [`AutoTokenizer`]:
```py
>>> from transformers import AutoTokenizer
>>> nome_del_modello = "nlptown/bert-base-multilingual-uncased-sentiment"
>>> tokenizer = AutoTokenizer.from_pretrained(nome_del_modello)
```
Dopodichรฉ, il tokenizer converte i token in numeri in modo da costruire un tensore come input del modello. Questo รจ conosciuto come il *vocabolario* del modello.
Passa il tuo testo al tokenizer:
```py
>>> encoding = tokenizer("Siamo molto felici di mostrarti la libreria ๐ค Transformers.")
>>> print(encoding)
{'input_ids': [101, 56821, 10132, 14407, 13019, 13007, 10120, 47201, 10330, 10106, 91686, 100, 58263, 119, 102],
'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}
```
Il tokenizer restituirร un dizionario contenente:
* [input_ids](./glossary#input-ids): rappresentazioni numeriche dei tuoi token.
* [attention_mask](.glossary#attention-mask): indica quali token devono essere presi in considerazione.
Come con la [`pipeline`], il tokenizer accetterร una lista di input. In piรน, il tokenizer puรฒ anche completare (pad, in inglese) e troncare il testo in modo da restituire un lotto (batch, in inglese) di lunghezza uniforme:
<frameworkcontent>
<pt>
```py
>>> pt_batch = tokenizer(
... ["Siamo molto felici di mostrarti la libreria ๐ค Transformers.", "Speriamo te non la odierai."],
... padding=True,
... truncation=True,
... max_length=512,
... return_tensors="pt",
... )
```
</pt>
<tf>
```py
>>> tf_batch = tokenizer(
... ["Siamo molto felici di mostrarti la libreria ๐ค Transformers.", "Speriamo te non la odierai."],
... padding=True,
... truncation=True,
... max_length=512,
... return_tensors="tf",
... )
```
</tf>
</frameworkcontent>
Leggi il tutorial sul [preprocessing](./preprocessing) per maggiori dettagli sulla tokenizzazione.
### AutoModel
<frameworkcontent>
<pt>
๐ค Transformers fornisce un metodo semplice e unificato per caricare istanze pre-allenate. Questo significa che puoi caricare un [`AutoModel`] come caricheresti un [`AutoTokenizer`]. L'unica differenza รจ selezionare l'[`AutoModel`] corretto per il compito di interesse. Dato che stai facendo classificazione di testi, o sequenze, carica [`AutoModelForSequenceClassification`]:
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model_name = "nlptown/bert-base-multilingual-uncased-sentiment"
>>> pt_model = AutoModelForSequenceClassification.from_pretrained(model_name)
```
<Tip>
Guarda il [task summary](./task_summary) per sapere quale classe di [`AutoModel`] utilizzare per quale compito.
</Tip>
Ora puoi passare il tuo lotto di input pre-processati direttamente al modello. Devi solo spacchettare il dizionario aggiungendo `**`:
```py
>>> pt_outputs = pt_model(**pt_batch)
```
Il modello produrrร le attivazioni finali nell'attributo `logits`. Applica la funzione softmax a `logits` per ottenere le probabilitร :
```py
>>> from torch import nn
>>> pt_predictions = nn.functional.softmax(pt_outputs.logits, dim=-1)
>>> print(pt_predictions)
tensor([[0.0041, 0.0037, 0.0203, 0.2005, 0.7713],
[0.3766, 0.3292, 0.1832, 0.0558, 0.0552]], grad_fn=<SoftmaxBackward0>)
```
</pt>
<tf>
๐ค Transformers fornisce un metodo semplice e unificato per caricare istanze pre-allenate. Questo significa che puoi caricare un [`TFAutoModel`] come caricheresti un [`AutoTokenizer`]. L'unica differenza รจ selezionare il [`TFAutoModel`] corretto per il compito di interesse. Dato che stai facendo classificazione di testi, o sequenze, carica [`TFAutoModelForSequenceClassification`]:
```py
>>> from transformers import TFAutoModelForSequenceClassification
>>> nome_del_modello = "nlptown/bert-base-multilingual-uncased-sentiment"
>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained(nome_del_modello)
```
<Tip>
Guarda il [task summary](./task_summary) per sapere quale classe di [`AutoModel`] utilizzare per quale compito.
</Tip>
Ora puoi passare il tuo lotto di input pre-processati direttamente al modello passando le chiavi del dizionario al tensore:
```py
>>> tf_outputs = tf_model(tf_batch)
```
Il modello produrrร le attivazioni finali nell'attributo `logits`. Applica la funzione softmax a `logits` per ottenere le probabilitร :
```py
>>> import tensorflow as tf
>>> tf_predictions = tf.nn.softmax(tf_outputs.logits, axis=-1)
>>> tf_predictions # doctest: +IGNORE_RESULT
```
</tf>
</frameworkcontent>
<Tip>
Tutti i modelli di ๐ค Transformers (PyTorch e TensorFlow) restituiscono i tensori *prima* della funzione finale
di attivazione (come la softmax) perchรฉ la funzione di attivazione finale viene spesso unita a quella di perdita.
</Tip>
I modelli sono [`torch.nn.Module`](https://pytorch.org/docs/stable/nn.html#torch.nn.Module) o [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) standard cosรฌ puoi utilizzarli all'interno del tuo training loop usuale. Tuttavia, per rendere le cose piรน semplici, ๐ค Transformers fornisce una classe [`Trainer`] per PyTorch che aggiunge delle funzionalitร per l'allenamento distribuito, precisione mista, e altro ancora. Per TensorFlow, puoi utilizzare il metodo `fit` di [Keras](https://keras.io/). Fai riferimento al [tutorial per il training](./training) per maggiori dettagli.
<Tip>
Gli output del modello di ๐ค Transformers sono delle dataclasses speciali in modo che i loro attributi vengano auto-completati all'interno di un IDE.
Gli output del modello si comportano anche come una tupla o un dizionario (ad esempio, puoi indicizzare con un intero, una slice o una stringa) nel qual caso gli attributi che sono `None` vengono ignorati.
</Tip>
### Salva un modello
<frameworkcontent>
<pt>
Una volta completato il fine-tuning del tuo modello, puoi salvarlo con il suo tokenizer utilizzando [`PreTrainedModel.save_pretrained`]:
```py
>>> pt_save_directory = "./pt_save_pretrained"
>>> tokenizer.save_pretrained(pt_save_directory) # doctest: +IGNORE_RESULT
>>> pt_model.save_pretrained(pt_save_directory)
```
Quando desideri utilizzare il tuo modello nuovamente, puoi ri-caricarlo con [`PreTrainedModel.from_pretrained`]:
```py
>>> pt_model = AutoModelForSequenceClassification.from_pretrained("./pt_save_pretrained")
```
</pt>
<tf>
Una volta completato il fine-tuning del tuo modello, puoi salvarlo con il suo tokenizer utilizzando [`TFPreTrainedModel.save_pretrained`]:
```py
>>> tf_save_directory = "./tf_save_pretrained"
>>> tokenizer.save_pretrained(tf_save_directory) # doctest: +IGNORE_RESULT
>>> tf_model.save_pretrained(tf_save_directory)
```
Quando desideri utilizzare il tuo modello nuovamente, puoi ri-caricarlo con [`TFPreTrainedModel.from_pretrained`]:
```py
>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained("./tf_save_pretrained")
```
</tf>
</frameworkcontent>
Una caratteristica particolarmente interessante di ๐ค Transformers รจ la sua abilitร di salvare un modello e ri-caricarlo sia come modello di PyTorch che di TensorFlow. I parametri `from_pt` o `from_tf` possono convertire un modello da un framework all'altro:
<frameworkcontent>
<pt>
```py
>>> from transformers import AutoModel
>>> tokenizer = AutoTokenizer.from_pretrained(tf_save_directory)
>>> pt_model = AutoModelForSequenceClassification.from_pretrained(tf_save_directory, from_tf=True)
```
</pt>
<tf>
```py
>>> from transformers import TFAutoModel
>>> tokenizer = AutoTokenizer.from_pretrained(pt_save_directory)
>>> tf_model = TFAutoModelForSequenceClassification.from_pretrained(pt_save_directory, from_pt=True)
```
</tf>
</frameworkcontent>
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/it/debugging.md
|
<!--Copyright 2021 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Debugging
## Debug dei problemi di rete multi-GPU
Quando addestri o fai inferenza con `DistributedDataParallel` e GPU multiple, se si verificano problemi di intercomunicazione tra processi e/o nodi, puoi utilizzare il seguente script per diagnosticare i problemi della rete.
```bash
wget https://raw.githubusercontent.com/huggingface/transformers/main/scripts/distributed/torch-distributed-gpu-test.py
```
Per esempio per testare come 2 GPU interagiscono fai:
```bash
python -m torch.distributed.run --nproc_per_node 2 --nnodes 1 torch-distributed-gpu-test.py
```
Se entrambi i processi sono in grado di comunicare tra loro e di allocare la memoria della GPU, ciascuno di essi stamperร lo stato OK.
Per piรน GPU o nodi adatta gli argumenti nello script.
All'interno dello script di diagnostica troverai molti altri dettagli e anche una guida per eseguirlo in ambiente SLURM.
Un livello di debug superiore รจ aggiungere la variabile d'ambiente `NCCL_DEBUG=INFO` come di seguito:
```bash
NCCL_DEBUG=INFO python -m torch.distributed.run --nproc_per_node 2 --nnodes 1 torch-distributed-gpu-test.py
```
In questo modo si scaricano molte informazioni di debug relative a NCCL, che puoi cercare online in caso di problemi. Oppure, se non hai la sicurezza di come interpretare l'output, puoi condividere il file di log in una Issue.
## Rilevamento di Underflow e Overflow
<Tip>
Questa funzionalitร al momento รจ disponibile solo per PyTorch.
</Tip>
<Tip>
Per addestramento multi-GPU richiede DDP (`torch.distributed.launch`).
</Tip>
<Tip>
Questa funzionalitร puรฒ essere usata con modelli basati su `nn.Module`.
</Tip>
Se inizi a ottenere `loss=NaN` o il modello presenta qualche altro comportamento anomalo a causa di valori `inf` o `nan` in
attivazioni o nei pesi, รจ necessario scoprire dove si verifica il primo underflow o overflow e cosa lo ha determinato. Fortunatamente
รจ possibile farlo facilmente attivando un modulo speciale che effettuerร il rilevamento automaticamente.
Se stai usando [`Trainer`], hai bisogno di aggiungere solo:
```bash
--debug underflow_overflow
```
ai normali argomenti della riga di comando, o passa `debug="underflow_overflow"` quando viene creato l'oggetto
[`TrainingArguments`].
Se stai usando il tuo ciclo di allenamento o un altro trainer, puoi ottenere lo stesso risultato con:
```python
from .debug_utils import DebugUnderflowOverflow
debug_overflow = DebugUnderflowOverflow(model)
```
[`~debug_utils.DebugUnderflowOverflow`] inserisce dei ganci nel modello che dopo ogni chiamata
testeranno le variabili di ingresso e di uscita e anche i pesi del modulo corrispondente. Non appena viene rilevato `inf` o
o `nan` in almeno un elemento delle attivazioni o dei pesi, il programma lo notifica e stampa un rapporto come il seguente (questo รจ stato rilevato con `google/mt5-small` sotto fp16 mixed precision):
```
Detected inf/nan during batch_number=0
Last 21 forward frames:
abs min abs max metadata
encoder.block.1.layer.1.DenseReluDense.dropout Dropout
0.00e+00 2.57e+02 input[0]
0.00e+00 2.85e+02 output
[...]
encoder.block.2.layer.0 T5LayerSelfAttention
6.78e-04 3.15e+03 input[0]
2.65e-04 3.42e+03 output[0]
None output[1]
2.25e-01 1.00e+04 output[2]
encoder.block.2.layer.1.layer_norm T5LayerNorm
8.69e-02 4.18e-01 weight
2.65e-04 3.42e+03 input[0]
1.79e-06 4.65e+00 output
encoder.block.2.layer.1.DenseReluDense.wi_0 Linear
2.17e-07 4.50e+00 weight
1.79e-06 4.65e+00 input[0]
2.68e-06 3.70e+01 output
encoder.block.2.layer.1.DenseReluDense.wi_1 Linear
8.08e-07 2.66e+01 weight
1.79e-06 4.65e+00 input[0]
1.27e-04 2.37e+02 output
encoder.block.2.layer.1.DenseReluDense.dropout Dropout
0.00e+00 8.76e+03 input[0]
0.00e+00 9.74e+03 output
encoder.block.2.layer.1.DenseReluDense.wo Linear
1.01e-06 6.44e+00 weight
0.00e+00 9.74e+03 input[0]
3.18e-04 6.27e+04 output
encoder.block.2.layer.1.DenseReluDense T5DenseGatedGeluDense
1.79e-06 4.65e+00 input[0]
3.18e-04 6.27e+04 output
encoder.block.2.layer.1.dropout Dropout
3.18e-04 6.27e+04 input[0]
0.00e+00 inf output
```
L'output di esempio รจ stato tagliato al centro per brevitร .
La seconda colonna mostra il valore dell'elemento piรน grande in assoluto,cosรฌ se osserviamo da vicino gli ultimi istanti,
input e output sono nel range di `1e4`. Questo addestramento รจ stato eseguito con una mixed precision fp16 e l'ultimo passo usciva fuori (sotto `fp16` il valore piรน grande prima di `inf` รจ `64e3`). Per evitare overflows sotto `fp16` le attivazionioni devono rimanere molto al di sotto di `1e4`, perchรฉ `1e4 * 1e4 = 1e8` quindi qualsiasi moltiplicazione di matrice con grandi attivazioni porterร a una condizione di overflow numerico.
All'inizio della traccia รจ possibile scoprire a quale lotto si รจ verificato il problema (questo `Detected inf/nan during batch_number=0` significa che il problema si รจ verificato nel primo lotto).
Ogni frame segnalato inizia dichiarando la voce completamente qualificata per il modulo corrispondente per il quale il frame รจ stato segnalato.
Se osserviamo il seguente frame:
```
encoder.block.2.layer.1.layer_norm T5LayerNorm
8.69e-02 4.18e-01 weight
2.65e-04 3.42e+03 input[0]
1.79e-06 4.65e+00 output
```
Questo, `encoder.block.2.layer.1.layer_norm` indica che si tratta di un layer norm nel primo layer, del secondo blocco dell'encoder. E le chiamata specifica di `forward` รจ `T5LayerNorm`.
Osserviamo gli ultimi frame del report:
```
Detected inf/nan during batch_number=0
Last 21 forward frames:
abs min abs max metadata
[...]
encoder.block.2.layer.1.DenseReluDense.wi_0 Linear
2.17e-07 4.50e+00 weight
1.79e-06 4.65e+00 input[0]
2.68e-06 3.70e+01 output
encoder.block.2.layer.1.DenseReluDense.wi_1 Linear
8.08e-07 2.66e+01 weight
1.79e-06 4.65e+00 input[0]
1.27e-04 2.37e+02 output
encoder.block.2.layer.1.DenseReluDense.wo Linear
1.01e-06 6.44e+00 weight
0.00e+00 9.74e+03 input[0]
3.18e-04 6.27e+04 output
encoder.block.2.layer.1.DenseReluDense T5DenseGatedGeluDense
1.79e-06 4.65e+00 input[0]
3.18e-04 6.27e+04 output
encoder.block.2.layer.1.dropout Dropout
3.18e-04 6.27e+04 input[0]
0.00e+00 inf output
```
L'ultimo frame report per la funzione `Dropout.forward` con la prima voce per l'unico input e la seconda per l'unico output. Si puรฒ notare che รจ stato richiamato da un attibuto `dropout` dentro la classe `DenseReluDense`. Si puรฒ notare che ciรฒ รจ avvenuto durante il primo strato, del 2ยฐ blocco, durante il primissimo lotto. Infine, gli elementi di input piรน grandi in assoluto sono stati `6.27e+04` e l'equivalente per l'output era `inf`.
Puoi vedere qui, che `T5DenseGatedGeluDense.forward` risulta in output activations, il cui valore massimo assoluto era circa 62,7K, che รจ molto vicino al limite massimo di 64K di fp16. Nel prossimo frame abbiamo `Dropout` che rinormalizza i pesi, dopo aver azzerato alcuni elementi, il che spinge il valore massimo assoluto a piรน di 64K e si verifica un overflow.(`inf`).
Come puoi notare, รจ nei frames precedenti che occorre esaminare quando i numeri iniziano a diventare molto grandi per i valori fp16.
Confrontiamo il report al codice `models/t5/modeling_t5.py`:
```python
class T5DenseGatedGeluDense(nn.Module):
def __init__(self, config):
super().__init__()
self.wi_0 = nn.Linear(config.d_model, config.d_ff, bias=False)
self.wi_1 = nn.Linear(config.d_model, config.d_ff, bias=False)
self.wo = nn.Linear(config.d_ff, config.d_model, bias=False)
self.dropout = nn.Dropout(config.dropout_rate)
self.gelu_act = ACT2FN["gelu_new"]
def forward(self, hidden_states):
hidden_gelu = self.gelu_act(self.wi_0(hidden_states))
hidden_linear = self.wi_1(hidden_states)
hidden_states = hidden_gelu * hidden_linear
hidden_states = self.dropout(hidden_states)
hidden_states = self.wo(hidden_states)
return hidden_states
```
Ora รจ facile vedere la chiamata `dropout`, e tutte le chiamate precedenti.
Poichรฉ il rilevamento avviene in un avanzamento (forward hook in eng.), i rapporti vengono creati immeditamente dopo ogni rientro da `forward` (forward returns in eng.).
Tornando al rapporto completo, per agire e risolvere il problema, dobbiamo andare qualche frame piรน in alto, dove i numeri hanno iniziato a salire, e probabilmente passare alla modalitร `fp32`, in modo che i numeri non trabocchino quando vengono moltiplicati o sommati. Naturalmente, potrebbero esserci altre soluzioni. Per esempio, potremmo spegnere temporanemante `amp` se รจ abilitato, successivamente spostare `forward` in un helper wrapper, come:
```python
def _forward(self, hidden_states):
hidden_gelu = self.gelu_act(self.wi_0(hidden_states))
hidden_linear = self.wi_1(hidden_states)
hidden_states = hidden_gelu * hidden_linear
hidden_states = self.dropout(hidden_states)
hidden_states = self.wo(hidden_states)
return hidden_states
import torch
def forward(self, hidden_states):
if torch.is_autocast_enabled():
with torch.cuda.amp.autocast(enabled=False):
return self._forward(hidden_states)
else:
return self._forward(hidden_states)
```
Poichรฉ il rilevatore automatico riporta solo gli ingressi e le uscite di fotogrammi completi, una volta che si sa dove cercare, si puรฒ
analizzare anche le fasi intermedie di una specifica funzione `forward`. In alcuni casi puoi usare la funzione di supporto `detect_overflow` per indirizzare il rilevatore dove preferisci, ad esempio:
```python
from debug_utils import detect_overflow
class T5LayerFF(nn.Module):
[...]
def forward(self, hidden_states):
forwarded_states = self.layer_norm(hidden_states)
detect_overflow(forwarded_states, "after layer_norm")
forwarded_states = self.DenseReluDense(forwarded_states)
detect_overflow(forwarded_states, "after DenseReluDense")
return hidden_states + self.dropout(forwarded_states)
```
Si puรฒ vedere che abbiamo aggiunto 2 di questi e ora teniamo traccia se `inf` o `nan` per `forwarded_states` รจ stato rilevato
da qualche parte.
In realtร , il rilevatore li riporta giร , perchรฉ ciascuna delle chiamate nell'esempio precedente รจ un `nn.Module`, ma
diciamo che se avessimo dei calcoli diretti locali, questo รจ il modo in cui lo faremmo.
Inoltre, se si istanzia il debugger nel proprio codice, รจ possibile modificare il numero di fotogrammi stampati rispetto a
predefinito, ad esempio.:
```python
from .debug_utils import DebugUnderflowOverflow
debug_overflow = DebugUnderflowOverflow(model, max_frames_to_save=100)
```
### Tracciamento della mistura assoluta del lotto specifico e del valore massimo
La stessa classe di debug puรฒ essere utilizzata per il tracciamento per-batch con la funzione di rilevamento di underflow/overflow disattivata.
Supponiamo di voler osservare i valori minimi e massimi assoluti per tutti gli ingredienti di ogni chiamata `forward` di un dato lotto.
lotto, e che lo si voglia fare solo per i lotti 1 e 3. Si istanzia questa classe come:
```python
debug_overflow = DebugUnderflowOverflow(model, trace_batch_nums=[1, 3])
```
Ora i batch completi 1 e 3 saranno tracciati utilizzando lo stesso formato del rilevatore di underflow/overflow.
I batches sono 0-indexed.
Questo รจ utile se si sa che il programma inizia a comportarsi male dopo un certo numero di batch, in modo da poter avanzare velocemente fino a quell'area.
direttamente a quell'area. Ecco un esempio di output troncato per questa configurazione:
```
*** Starting batch number=1 ***
abs min abs max metadata
shared Embedding
1.01e-06 7.92e+02 weight
0.00e+00 2.47e+04 input[0]
5.36e-05 7.92e+02 output
[...]
decoder.dropout Dropout
1.60e-07 2.27e+01 input[0]
0.00e+00 2.52e+01 output
decoder T5Stack
not a tensor output
lm_head Linear
1.01e-06 7.92e+02 weight
0.00e+00 1.11e+00 input[0]
6.06e-02 8.39e+01 output
T5ForConditionalGeneration
not a tensor output
*** Starting batch number=3 ***
abs min abs max metadata
shared Embedding
1.01e-06 7.92e+02 weight
0.00e+00 2.78e+04 input[0]
5.36e-05 7.92e+02 output
[...]
```
Qui verrร scaricato un numero enorme di fotogrammi, tanti quanti sono le chiamate in avanti nel modello, quindi puรฒ essere o non essere quello che volete, ma a volte puรฒ essere piรน utile usarlo di un classico debugger. Per esempio, se il problema inizia a verificarsi a partire dal lotto numero 150. Quindi รจ possibile scaricare le tracce dei lotti 149 e 150 e confrontare i punti in cui i numeri hanno iniziato a divergere.
ร inoltre possibile specificare il numero di batch dopo il quale interrompere l'addestramento, con:
```python
debug_overflow = DebugUnderflowOverflow(model, trace_batch_nums=[1, 3], abort_after_batch_num=3)
```
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/multilingual.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ๋ค๊ตญ์ด ๋ชจ๋ธ ์ถ๋ก ํ๊ธฐ[[multilingual-models-for-inference]]
[[open-in-colab]]
๐ค Transformers์๋ ์ฌ๋ฌ ์ข
๋ฅ์ ๋ค๊ตญ์ด(multilingual) ๋ชจ๋ธ์ด ์์ผ๋ฉฐ, ๋จ์ผ ์ธ์ด(monolingual) ๋ชจ๋ธ๊ณผ ์ถ๋ก ์ ์ฌ์ฉ๋ฒ์ด ๋ค๋ฆ
๋๋ค.
๊ทธ๋ ๋ค๊ณ ํด์ *๋ชจ๋ * ๋ค๊ตญ์ด ๋ชจ๋ธ์ ์ฌ์ฉ๋ฒ์ด ๋ค๋ฅธ ๊ฒ์ ์๋๋๋ค.
[bert-base-multilingual-uncased](https://huggingface.co/bert-base-multilingual-uncased)์ ๊ฐ์ ๋ช๋ช ๋ชจ๋ธ์ ๋จ์ผ ์ธ์ด ๋ชจ๋ธ์ฒ๋ผ ์ฌ์ฉํ ์ ์์ต๋๋ค.
์ด๋ฒ ๊ฐ์ด๋์์ ๋ค๊ตญ์ด ๋ชจ๋ธ์ ์ถ๋ก ์ ์ฌ์ฉ ๋ฐฉ๋ฒ์ ์์๋ณผ ๊ฒ์
๋๋ค.
## XLM[[xlm]]
XLM์๋ 10๊ฐ์ง ์ฒดํฌํฌ์ธํธ(checkpoint)๊ฐ ์๋๋ฐ, ์ด ์ค ํ๋๋ง ๋จ์ผ ์ธ์ด์
๋๋ค.
๋๋จธ์ง ์ฒดํฌํฌ์ธํธ 9๊ฐ๋ ์ธ์ด ์๋ฒ ๋ฉ์ ์ฌ์ฉํ๋ ์ฒดํฌํฌ์ธํธ์ ๊ทธ๋ ์ง ์์ ์ฒดํฌํฌ์ธํธ์ ๋ ๊ฐ์ง ๋ฒ์ฃผ๋ก ๋๋ ์ ์์ต๋๋ค.
### ์ธ์ด ์๋ฒ ๋ฉ์ ์ฌ์ฉํ๋ XLM[[xlm-with-language-embeddings]]
๋ค์ XLM ๋ชจ๋ธ์ ์ถ๋ก ์์ ์ธ์ด ์๋ฒ ๋ฉ์ ์ฌ์ฉํฉ๋๋ค:
- `xlm-mlm-ende-1024` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง, ์์ด-๋
์ผ์ด)
- `xlm-mlm-enfr-1024` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง, ์์ด-ํ๋์ค์ด)
- `xlm-mlm-enro-1024` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง, ์์ด-๋ฃจ๋ง๋์์ด)
- `xlm-mlm-xnli15-1024` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง, XNLI ๋ฐ์ดํฐ ์ธํธ์์ ์ ๊ณตํ๋ 15๊ฐ ๊ตญ์ด)
- `xlm-mlm-tlm-xnli15-1024` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง + ๋ฒ์ญ, XNLI ๋ฐ์ดํฐ ์ธํธ์์ ์ ๊ณตํ๋ 15๊ฐ ๊ตญ์ด)
- `xlm-clm-enfr-1024` (Causal language modeling, ์์ด-ํ๋์ค์ด)
- `xlm-clm-ende-1024` (Causal language modeling, ์์ด-๋
์ผ์ด)
์ธ์ด ์๋ฒ ๋ฉ์ ๋ชจ๋ธ์ ์ ๋ฌ๋ `input_ids`์ ๋์ผํ shape์ ํ
์๋ก ํํ๋ฉ๋๋ค.
์ด๋ฌํ ํ
์์ ๊ฐ์ ์ฌ์ฉ๋ ์ธ์ด์ ๋ฐ๋ผ ๋ค๋ฅด๋ฉฐ ํ ํฌ๋์ด์ ์ `lang2id` ๋ฐ `id2lang` ์์ฑ์ ์ํด ์๋ณ๋ฉ๋๋ค.
๋ค์ ์์ ์์๋ `xlm-clm-enfr-1024` ์ฒดํฌํฌ์ธํธ(์ฝ์ ์ธ์ด ๋ชจ๋ธ๋ง(causal language modeling), ์์ด-ํ๋์ค์ด)๋ฅผ ๊ฐ์ ธ์ต๋๋ค:
```py
>>> import torch
>>> from transformers import XLMTokenizer, XLMWithLMHeadModel
>>> tokenizer = XLMTokenizer.from_pretrained("xlm-clm-enfr-1024")
>>> model = XLMWithLMHeadModel.from_pretrained("xlm-clm-enfr-1024")
```
ํ ํฌ๋์ด์ ์ `lang2id` ์์ฑ์ ๋ชจ๋ธ์ ์ธ์ด์ ํด๋น ID๋ฅผ ํ์ํฉ๋๋ค:
```py
>>> print(tokenizer.lang2id)
{'en': 0, 'fr': 1}
```
๋ค์์ผ๋ก, ์์ ์
๋ ฅ์ ๋ง๋ญ๋๋ค:
```py
>>> input_ids = torch.tensor([tokenizer.encode("Wikipedia was used to")]) # ๋ฐฐ์น ํฌ๊ธฐ๋ 1์
๋๋ค
```
์ธ์ด ID๋ฅผ `"en"`์ผ๋ก ์ค์ ํด ์ธ์ด ์๋ฒ ๋ฉ์ ์ ์ํฉ๋๋ค.
์ธ์ด ์๋ฒ ๋ฉ์ ์์ด์ ์ธ์ด ID์ธ `0`์ผ๋ก ์ฑ์์ง ํ
์์
๋๋ค.
์ด ํ
์๋ `input_ids`์ ๊ฐ์ ํฌ๊ธฐ์ฌ์ผ ํฉ๋๋ค.
```py
>>> language_id = tokenizer.lang2id["en"] # 0
>>> langs = torch.tensor([language_id] * input_ids.shape[1]) # torch.tensor([0, 0, 0, ..., 0])
>>> # (batch_size, sequence_length) shape์ ํ
์๊ฐ ๋๋๋ก ๋ง๋ญ๋๋ค.
>>> langs = langs.view(1, -1) # ์ด์ [1, sequence_length] shape์ด ๋์์ต๋๋ค(๋ฐฐ์น ํฌ๊ธฐ๋ 1์
๋๋ค)
```
์ด์ `input_ids`์ ์ธ์ด ์๋ฒ ๋ฉ์ ๋ชจ๋ธ๋ก ์ ๋ฌํฉ๋๋ค:
```py
>>> outputs = model(input_ids, langs=langs)
```
[run_generation.py](https://github.com/huggingface/transformers/tree/main/examples/pytorch/text-generation/run_generation.py) ์คํฌ๋ฆฝํธ๋ก `xlm-clm` ์ฒดํฌํฌ์ธํธ๋ฅผ ์ฌ์ฉํด ํ
์คํธ์ ์ธ์ด ์๋ฒ ๋ฉ์ ์์ฑํ ์ ์์ต๋๋ค.
### ์ธ์ด ์๋ฒ ๋ฉ์ ์ฌ์ฉํ์ง ์๋ XLM[[xlm-without-language-embeddings]]
๋ค์ XLM ๋ชจ๋ธ์ ์ถ๋ก ์์ ์ธ์ด ์๋ฒ ๋ฉ์ด ํ์ํ์ง ์์ต๋๋ค:
- `xlm-mlm-17-1280` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง, 17๊ฐ ๊ตญ์ด)
- `xlm-mlm-100-1280` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง, 100๊ฐ ๊ตญ์ด)
์ด์ ์ XLM ์ฒดํฌํฌ์ธํธ์ ๋ฌ๋ฆฌ ์ด ๋ชจ๋ธ์ ์ผ๋ฐ ๋ฌธ์ฅ ํํ์ ์ฌ์ฉ๋ฉ๋๋ค.
## BERT[[bert]]
๋ค์ BERT ๋ชจ๋ธ์ ๋ค๊ตญ์ด ํ์คํฌ์ ์ฌ์ฉํ ์ ์์ต๋๋ค:
- `bert-base-multilingual-uncased` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง + ๋ค์ ๋ฌธ์ฅ ์์ธก, 102๊ฐ ๊ตญ์ด)
- `bert-base-multilingual-cased` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง + ๋ค์ ๋ฌธ์ฅ ์์ธก, 104๊ฐ ๊ตญ์ด)
์ด๋ฌํ ๋ชจ๋ธ์ ์ถ๋ก ์์ ์ธ์ด ์๋ฒ ๋ฉ์ด ํ์ํ์ง ์์ต๋๋ค.
๋ฌธ๋งฅ์์ ์ธ์ด๋ฅผ ์๋ณํ๊ณ , ์๋ณ๋ ์ธ์ด๋ก ์ถ๋ก ํฉ๋๋ค.
## XLM-RoBERTa[[xlmroberta]]
๋ค์ XLM-RoBERTa ๋ํ ๋ค๊ตญ์ด ๋ค๊ตญ์ด ํ์คํฌ์ ์ฌ์ฉํ ์ ์์ต๋๋ค:
- `xlm-roberta-base` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง, 100๊ฐ ๊ตญ์ด)
- `xlm-roberta-large` (๋ง์คํน๋ ์ธ์ด ๋ชจ๋ธ๋ง, 100๊ฐ ๊ตญ์ด)
XLM-RoBERTa๋ 100๊ฐ ๊ตญ์ด์ ๋ํด ์๋ก ์์ฑ๋๊ณ ์ ์ ๋ 2.5TB ๊ท๋ชจ์ CommonCrawl ๋ฐ์ดํฐ๋ก ํ์ต๋์์ต๋๋ค.
์ด์ ์ ๊ณต๊ฐ๋ mBERT๋ XLM๊ณผ ๊ฐ์ ๋ค๊ตญ์ด ๋ชจ๋ธ์ ๋นํด ๋ถ๋ฅ, ์ํ์ค ๋ผ๋ฒจ๋ง, ์ง์ ์๋ต๊ณผ ๊ฐ์ ๋ค์ด์คํธ๋ฆผ(downstream) ์์
์์ ์ด์ ์ด ์์ต๋๋ค.
## M2M100[[m2m100]]
๋ค์ M2M100 ๋ชจ๋ธ ๋ํ ๋ค๊ตญ์ด ๋ค๊ตญ์ด ํ์คํฌ์ ์ฌ์ฉํ ์ ์์ต๋๋ค:
- `facebook/m2m100_418M` (๋ฒ์ญ)
- `facebook/m2m100_1.2B` (๋ฒ์ญ)
์ด ์์ ์์๋ `facebook/m2m100_418M` ์ฒดํฌํฌ์ธํธ๋ฅผ ๊ฐ์ ธ์์ ์ค๊ตญ์ด๋ฅผ ์์ด๋ก ๋ฒ์ญํฉ๋๋ค.
ํ ํฌ๋์ด์ ์์ ๋ฒ์ญ ๋์ ์ธ์ด(source language)๋ฅผ ์ค์ ํ ์ ์์ต๋๋ค:
```py
>>> from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer
>>> en_text = "Do not meddle in the affairs of wizards, for they are subtle and quick to anger."
>>> chinese_text = "ไธ่ฆๆๆๅทซๅธซ็ไบๅ, ๅ ็บไปๅๆฏๅพฎๅฆ็, ๅพๅฟซๅฐฑๆ็ผๆ."
>>> tokenizer = M2M100Tokenizer.from_pretrained("facebook/m2m100_418M", src_lang="zh")
>>> model = M2M100ForConditionalGeneration.from_pretrained("facebook/m2m100_418M")
```
๋ฌธ์ฅ์ ํ ํฐํํฉ๋๋ค:
```py
>>> encoded_zh = tokenizer(chinese_text, return_tensors="pt")
```
M2M100์ ๋ฒ์ญ์ ์งํํ๊ธฐ ์ํด ์ฒซ ๋ฒ์งธ๋ก ์์ฑ๋๋ ํ ํฐ์ ๋ฒ์ญํ ์ธ์ด(target language) ID๋ก ๊ฐ์ ์ง์ ํฉ๋๋ค.
์์ด๋ก ๋ฒ์ญํ๊ธฐ ์ํด `generate` ๋ฉ์๋์์ `forced_bos_token_id`๋ฅผ `en`์ผ๋ก ์ค์ ํฉ๋๋ค:
```py
>>> generated_tokens = model.generate(**encoded_zh, forced_bos_token_id=tokenizer.get_lang_id("en"))
>>> tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
'Do not interfere with the matters of the witches, because they are delicate and will soon be angry.'
```
## MBart[[mbart]]
๋ค์ MBart ๋ชจ๋ธ ๋ํ ๋ค๊ตญ์ด ํ์คํฌ์ ์ฌ์ฉํ ์ ์์ต๋๋ค:
- `facebook/mbart-large-50-one-to-many-mmt` (์ผ๋๋ค ๋ค๊ตญ์ด ๋ฒ์ญ, 50๊ฐ ๊ตญ์ด)
- `facebook/mbart-large-50-many-to-many-mmt` (๋ค๋๋ค ๋ค๊ตญ์ด ๋ฒ์ญ, 50๊ฐ ๊ตญ์ด)
- `facebook/mbart-large-50-many-to-one-mmt` (๋ค๋์ผ ๋ค๊ตญ์ด ๋ฒ์ญ, 50๊ฐ ๊ตญ์ด)
- `facebook/mbart-large-50` (๋ค๊ตญ์ด ๋ฒ์ญ, 50๊ฐ ๊ตญ์ด)
- `facebook/mbart-large-cc25`
์ด ์์ ์์๋ ํ๋๋์ด๋ฅผ ์์ด๋ก ๋ฒ์ญํ๊ธฐ ์ํด `facebook/mbart-large-50-many-to-many-mmt` ์ฒดํฌํฌ์ธํธ๋ฅผ ๊ฐ์ ธ์ต๋๋ค.
ํ ํฌ๋์ด์ ์์ ๋ฒ์ญ ๋์ ์ธ์ด(source language)๋ฅผ ์ค์ ํ ์ ์์ต๋๋ค:
```py
>>> from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
>>> en_text = "Do not meddle in the affairs of wizards, for they are subtle and quick to anger."
>>> fi_text = "รlรค sekaannu velhojen asioihin, sillรค ne ovat hienovaraisia ja nopeasti vihaisia."
>>> tokenizer = AutoTokenizer.from_pretrained("facebook/mbart-large-50-many-to-many-mmt", src_lang="fi_FI")
>>> model = AutoModelForSeq2SeqLM.from_pretrained("facebook/mbart-large-50-many-to-many-mmt")
```
๋ฌธ์ฅ์ ํ ํฐํํฉ๋๋ค:
```py
>>> encoded_en = tokenizer(en_text, return_tensors="pt")
```
MBart๋ ๋ฒ์ญ์ ์งํํ๊ธฐ ์ํด ์ฒซ ๋ฒ์งธ๋ก ์์ฑ๋๋ ํ ํฐ์ ๋ฒ์ญํ ์ธ์ด(target language) ID๋ก ๊ฐ์ ์ง์ ํฉ๋๋ค.
์์ด๋ก ๋ฒ์ญํ๊ธฐ ์ํด `generate` ๋ฉ์๋์์ `forced_bos_token_id`๋ฅผ `en`์ผ๋ก ์ค์ ํฉ๋๋ค:
```py
>>> generated_tokens = model.generate(**encoded_en, forced_bos_token_id=tokenizer.lang_code_to_id("en_XX"))
>>> tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
"Don't interfere with the wizard's affairs, because they are subtle, will soon get angry."
```
`facebook/mbart-large-50-many-to-one-mmt` ์ฒดํฌํฌ์ธํธ๋ฅผ ์ฌ์ฉํ๊ณ ์๋ค๋ฉด, ์ฒซ ๋ฒ์งธ๋ก ์์ฑ๋๋ ํ ํฐ์ ๋ฒ์ญํ ์ธ์ด(target language) ID๋ก ๊ฐ์ ์ง์ ํ ํ์๋ ์์ต๋๋ค.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/tflite.md
|
<!--Copyright 2023 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# TFLite๋ก ๋ด๋ณด๋ด๊ธฐ[[export-to-tflite]]
[TensorFlow Lite](https://www.tensorflow.org/lite/guide)๋ ์์์ด ์ ํ๋ ํด๋ํฐ, ์๋ฒ ๋๋ ์์คํ
, ์ฌ๋ฌผ์ธํฐ๋ท(IoT) ๊ธฐ๊ธฐ์์
๊ธฐ๊ณํ์ต ๋ชจ๋ธ์ ๋ฐฐํฌํ๊ธฐ ์ํ ๊ฒฝ๋ ํ๋ ์์ํฌ์
๋๋ค.
TFLite๋ ์ฐ์ฐ ๋ฅ๋ ฅ, ๋ฉ๋ชจ๋ฆฌ, ์ ๋ ฅ ์๋น๊ฐ ์ ํ๋ ๊ธฐ๊ธฐ์์ ๋ชจ๋ธ์ ํจ์จ์ ์ผ๋ก ์ต์ ํํ๊ณ ์คํํ๊ธฐ ์ํด
์ค๊ณ๋์์ต๋๋ค.
TensorFlow Lite ๋ชจ๋ธ์ `.tflite` ํ์ผ ํ์ฅ์๋ก ์๋ณ๋๋ ํน์ํ๊ณ ํจ์จ์ ์ธ ํด๋์ฉ ํฌ๋งท์ผ๋ก ํํ๋ฉ๋๋ค.
๐ค Optimum์ `exporters.tflite` ๋ชจ๋๋ก ๐ค Transformers ๋ชจ๋ธ์ TFLite๋ก ๋ด๋ณด๋ด๋ ๊ธฐ๋ฅ์ ์ ๊ณตํฉ๋๋ค.
์ง์๋๋ ๋ชจ๋ธ ์ํคํ
์ฒ ๋ชฉ๋ก์ [๐ค Optimum ๋ฌธ์](https://huggingface.co/docs/optimum/exporters/tflite/overview)๋ฅผ ์ฐธ๊ณ ํ์ธ์.
๋ชจ๋ธ์ TFLite๋ก ๋ด๋ณด๋ด๋ ค๋ฉด, ํ์ํ ์ข
์์ฑ์ ์ค์นํ์ธ์:
```bash
pip install optimum[exporters-tf]
```
๋ชจ๋ ์ฌ์ฉ ๊ฐ๋ฅํ ์ธ์๋ฅผ ํ์ธํ๋ ค๋ฉด, [๐ค Optimum ๋ฌธ์](https://huggingface.co/docs/optimum/main/en/exporters/tflite/usage_guides/export_a_model)๋ฅผ ์ฐธ๊ณ ํ๊ฑฐ๋
ํฐ๋ฏธ๋์์ ๋์๋ง์ ์ดํด๋ณด์ธ์:
```bash
optimum-cli export tflite --help
```
์๋ฅผ ๋ค์ด ๐ค Hub์์์ `bert-base-uncased` ๋ชจ๋ธ ์ฒดํฌํฌ์ธํธ๋ฅผ ๋ด๋ณด๋ด๋ ค๋ฉด, ๋ค์ ๋ช
๋ น์ ์คํํ์ธ์:
```bash
optimum-cli export tflite --model bert-base-uncased --sequence_length 128 bert_tflite/
```
๋ค์๊ณผ ๊ฐ์ด ์งํ ์ํฉ์ ๋ํ๋ด๋ ๋ก๊ทธ์ ๊ฒฐ๊ณผ๋ฌผ์ธ `model.tflite`๊ฐ ์ ์ฅ๋ ์์น๋ฅผ ๋ณด์ฌ์ฃผ๋ ๋ก๊ทธ๊ฐ ํ์๋ฉ๋๋ค:
```bash
Validating TFLite model...
-[โ] TFLite model output names match reference model (logits)
- Validating TFLite Model output "logits":
-[โ] (1, 128, 30522) matches (1, 128, 30522)
-[x] values not close enough, max diff: 5.817413330078125e-05 (atol: 1e-05)
The TensorFlow Lite export succeeded with the warning: The maximum absolute difference between the output of the reference model and the TFLite exported model is not within the set tolerance 1e-05:
- logits: max diff = 5.817413330078125e-05.
The exported model was saved at: bert_tflite
```
์ ์์ ๋ ๐ค Hub์์์ ์ฒดํฌํฌ์ธํธ๋ฅผ ๋ด๋ณด๋ด๋ ๋ฐฉ๋ฒ์ ๋ณด์ฌ์ค๋๋ค.
๋ก์ปฌ ๋ชจ๋ธ์ ๋ด๋ณด๋ธ๋ค๋ฉด, ๋จผ์ ๋ชจ๋ธ ๊ฐ์ค์น์ ํ ํฌ๋์ด์ ํ์ผ์ด ๋ชจ๋ ๊ฐ์ ๋๋ ํฐ๋ฆฌ( `local_path` )์ ์ ์ฅ๋๋์ง ํ์ธํ์ธ์.
CLI๋ฅผ ์ฌ์ฉํ ๋, ๐ค Hub์์์ ์ฒดํฌํฌ์ธํธ ์ด๋ฆ ๋์ `model` ์ธ์์ `local_path`๋ฅผ ์ ๋ฌํ๋ฉด ๋ฉ๋๋ค.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/preprocessing.md
|
<!--Copyright 2023 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ์ ์ฒ๋ฆฌ[[preprocess]]
[[open-in-colab]]
๋ชจ๋ธ์ ํ๋ จํ๋ ค๋ฉด ๋ฐ์ดํฐ ์ธํธ๋ฅผ ๋ชจ๋ธ์ ๋ง๋ ์
๋ ฅ ํ์์ผ๋ก ์ ์ฒ๋ฆฌํด์ผ ํฉ๋๋ค. ํ
์คํธ, ์ด๋ฏธ์ง ๋๋ ์ค๋์ค์ธ์ง ๊ด๊ณ์์ด ๋ฐ์ดํฐ๋ฅผ ํ
์ ๋ฐฐ์น๋ก ๋ณํํ๊ณ ์กฐ๋ฆฝํ ํ์๊ฐ ์์ต๋๋ค. ๐ค Transformers๋ ๋ชจ๋ธ์ ๋ํ ๋ฐ์ดํฐ๋ฅผ ์ค๋นํ๋ ๋ฐ ๋์์ด ๋๋ ์ผ๋ จ์ ์ ์ฒ๋ฆฌ ํด๋์ค๋ฅผ ์ ๊ณตํฉ๋๋ค. ์ด ํํ ๋ฆฌ์ผ์์๋ ๋ค์ ๋ด์ฉ์ ๋ฐฐ์ธ ์ ์์ต๋๋ค:
* ํ
์คํธ๋ [Tokenizer](./main_classes/tokenizer)๋ฅผ ์ฌ์ฉํ์ฌ ํ ํฐ ์ํ์ค๋ก ๋ณํํ๊ณ ํ ํฐ์ ์ซ์ ํํ์ ๋ง๋ ํ ํ
์๋ก ์กฐ๋ฆฝํฉ๋๋ค.
* ์์ฑ ๋ฐ ์ค๋์ค๋ [Feature extractor](./main_classes/feature_extractor)๋ฅผ ์ฌ์ฉํ์ฌ ์ค๋์ค ํํ์์ ์ํ์ค ํน์ฑ์ ํ์
ํ์ฌ ํ
์๋ก ๋ณํํฉ๋๋ค.
* ์ด๋ฏธ์ง ์
๋ ฅ์ [ImageProcessor](./main_classes/image)์ ์ฌ์ฉํ์ฌ ์ด๋ฏธ์ง๋ฅผ ํ
์๋ก ๋ณํํฉ๋๋ค.
* ๋ฉํฐ๋ชจ๋ฌ ์
๋ ฅ์ [Processor](./main_classes/processors)์ ์ฌ์ฉํ์ฌ ํ ํฌ๋์ด์ ์ ํน์ฑ ์ถ์ถ๊ธฐ ๋๋ ์ด๋ฏธ์ง ํ๋ก์ธ์๋ฅผ ๊ฒฐํฉํฉ๋๋ค.
<Tip>
`AutoProcessor`๋ **์ธ์ ๋** ์๋ํ์ฌ ํ ํฌ๋์ด์ , ์ด๋ฏธ์ง ํ๋ก์ธ์, ํน์ฑ ์ถ์ถ๊ธฐ ๋๋ ํ๋ก์ธ์ ๋ฑ ์ฌ์ฉ ์ค์ธ ๋ชจ๋ธ์ ๋ง๋ ํด๋์ค๋ฅผ ์๋์ผ๋ก ์ ํํฉ๋๋ค.
</Tip>
์์ํ๊ธฐ ์ ์ ๐ค Datasets๋ฅผ ์ค์นํ์ฌ ์คํ์ ์ฌ์ฉํ ๋ฐ์ดํฐ๋ฅผ ๋ถ๋ฌ์ฌ ์ ์์ต๋๋ค:
```bash
pip install datasets
```
## ์์ฐ์ด์ฒ๋ฆฌ[[natural-language-processing]]
<Youtube id="Yffk5aydLzg"/>
ํ
์คํธ ๋ฐ์ดํฐ๋ฅผ ์ ์ฒ๋ฆฌํ๊ธฐ ์ํ ๊ธฐ๋ณธ ๋๊ตฌ๋ [tokenizer](main_classes/tokenizer)์
๋๋ค. ํ ํฌ๋์ด์ ๋ ์ผ๋ จ์ ๊ท์น์ ๋ฐ๋ผ ํ
์คํธ๋ฅผ *ํ ํฐ*์ผ๋ก ๋๋๋๋ค. ํ ํฐ์ ์ซ์๋ก ๋ณํ๋๊ณ ํ
์๋ ๋ชจ๋ธ ์
๋ ฅ์ด ๋ฉ๋๋ค. ๋ชจ๋ธ์ ํ์ํ ์ถ๊ฐ ์
๋ ฅ์ ํ ํฌ๋์ด์ ์ ์ํด ์ถ๊ฐ๋ฉ๋๋ค.
<Tip>
์ฌ์ ํ๋ จ๋ ๋ชจ๋ธ์ ์ฌ์ฉํ ๊ณํ์ด๋ผ๋ฉด ๋ชจ๋ธ๊ณผ ํจ๊ป ์ฌ์ ํ๋ จ๋ ํ ํฌ๋์ด์ ๋ฅผ ์ฌ์ฉํ๋ ๊ฒ์ด ์ค์ํฉ๋๋ค. ์ด๋ ๊ฒ ํ๋ฉด ํ
์คํธ๊ฐ ์ฌ์ ํ๋ จ ๋ง๋ญ์น์ ๋์ผํ ๋ฐฉ์์ผ๋ก ๋ถํ ๋๊ณ ์ฌ์ ํ๋ จ ์ค์ ๋์ผํ ํด๋น ํ ํฐ-์ธ๋ฑ์ค ์(์ผ๋ฐ์ ์ผ๋ก *vocab*์ด๋ผ๊ณ ํจ)์ ์ฌ์ฉํฉ๋๋ค.
</Tip>
์์ํ๋ ค๋ฉด [`AutoTokenizer.from_pretrained`] ๋ฉ์๋๋ฅผ ์ฌ์ฉํ์ฌ ์ฌ์ ํ๋ จ๋ ํ ํฌ๋์ด์ ๋ฅผ ๋ถ๋ฌ์ค์ธ์. ๋ชจ๋ธ๊ณผ ํจ๊ป ์ฌ์ ํ๋ จ๋ *vocab*์ ๋ค์ด๋ก๋ํฉ๋๋ค:
```py
>>> from transformers import AutoTokenizer
>>> tokenizer = AutoTokenizer.from_pretrained("bert-base-cased")
```
๊ทธ ๋ค์์ผ๋ก ํ
์คํธ๋ฅผ ํ ํฌ๋์ด์ ์ ๋ฃ์ด์ฃผ์ธ์:
```py
>>> encoded_input = tokenizer("Do not meddle in the affairs of wizards, for they are subtle and quick to anger.")
>>> print(encoded_input)
{'input_ids': [101, 2079, 2025, 19960, 10362, 1999, 1996, 3821, 1997, 16657, 1010, 2005, 2027, 2024, 11259, 1998, 4248, 2000, 4963, 1012, 102],
'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]}
```
ํ ํฌ๋์ด์ ๋ ์ธ ๊ฐ์ง ์ค์ํ ํญ๋ชฉ์ ํฌํจํ ๋์
๋๋ฆฌ๋ฅผ ๋ฐํํฉ๋๋ค:
* [input_ids](glossary#input-ids)๋ ๋ฌธ์ฅ์ ๊ฐ ํ ํฐ์ ํด๋นํ๋ ์ธ๋ฑ์ค์
๋๋ค.
* [attention_mask](glossary#attention-mask)๋ ํ ํฐ์ ์ฒ๋ฆฌํด์ผ ํ๋์ง ์ฌ๋ถ๋ฅผ ๋ํ๋
๋๋ค.
* [token_type_ids](glossary#token-type-ids)๋ ๋ ๊ฐ ์ด์์ ์ํ์ค๊ฐ ์์ ๋ ํ ํฐ์ด ์ํ ์ํ์ค๋ฅผ ์๋ณํฉ๋๋ค.
`input_ids`๋ฅผ ๋์ฝ๋ฉํ์ฌ ์
๋ ฅ์ ๋ฐํํฉ๋๋ค:
```py
>>> tokenizer.decode(encoded_input["input_ids"])
'[CLS] Do not meddle in the affairs of wizards, for they are subtle and quick to anger. [SEP]'
```
ํ ํฌ๋์ด์ ๊ฐ ๋ ๊ฐ์ ํน์ํ ํ ํฐ(๋ถ๋ฅ ํ ํฐ `CLS`์ ๋ถํ ํ ํฐ `SEP`)์ ๋ฌธ์ฅ์ ์ถ๊ฐํ์ต๋๋ค.
๋ชจ๋ ๋ชจ๋ธ์ ํน์ํ ํ ํฐ์ด ํ์ํ ๊ฒ์ ์๋์ง๋ง, ํ์ํ๋ค๋ฉด ํ ํฌ๋์ด์ ๊ฐ ์๋์ผ๋ก ์ถ๊ฐํฉ๋๋ค.
์ ์ฒ๋ฆฌํ ๋ฌธ์ฅ์ด ์ฌ๋ฌ ๊ฐ ์๋ ๊ฒฝ์ฐ์๋ ๋ฆฌ์คํธ๋ก ํ ํฌ๋์ด์ ์ ์ ๋ฌํฉ๋๋ค:
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_inputs = tokenizer(batch_sentences)
>>> print(encoded_inputs)
{'input_ids': [[101, 1252, 1184, 1164, 1248, 6462, 136, 102],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102]],
'token_type_ids': [[0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0]],
'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1]]}
```
### ํจ๋ฉ[[pad]]
๋ชจ๋ธ ์
๋ ฅ์ธ ํ
์๋ ๋ชจ์์ด ๊ท ์ผํด์ผ ํ์ง๋ง, ๋ฌธ์ฅ์ ๊ธธ์ด๊ฐ ํญ์ ๊ฐ์ง๋ ์๊ธฐ ๋๋ฌธ์ ๋ฌธ์ ๊ฐ ๋ ์ ์์ต๋๋ค. ํจ๋ฉ์ ์งง์ ๋ฌธ์ฅ์ ํน์ํ *ํจ๋ฉ ํ ํฐ*์ ์ถ๊ฐํ์ฌ ํ
์๋ฅผ ์ง์ฌ๊ฐํ ๋ชจ์์ด ๋๋๋ก ํ๋ ์ ๋ต์
๋๋ค.
`padding` ๋งค๊ฐ๋ณ์๋ฅผ `True`๋ก ์ค์ ํ์ฌ ๋ฐฐ์น ๋ด์ ์งง์ ์ํ์ค๋ฅผ ๊ฐ์ฅ ๊ธด ์ํ์ค์ ๋ง์ถฐ ํจ๋ฉํฉ๋๋ค.
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch_sentences, padding=True)
>>> print(encoded_input)
{'input_ids': [[101, 1252, 1184, 1164, 1248, 6462, 136, 102, 0, 0, 0, 0, 0, 0, 0],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102, 0, 0, 0, 0, 0, 0, 0, 0]],
'token_type_ids': [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]],
'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]]}
```
๊ธธ์ด๊ฐ ์งง์ ์ฒซ ๋ฌธ์ฅ๊ณผ ์ธ ๋ฒ์งธ ๋ฌธ์ฅ์ด ์ด์ `0`์ผ๋ก ์ฑ์์ก์ต๋๋ค.
### ์๋ผ๋ด๊ธฐ[[truncation]]
ํํธ, ๋๋ก๋ ์ํ์ค๊ฐ ๋ชจ๋ธ์์ ์ฒ๋ฆฌํ๊ธฐ์ ๋๋ฌด ๊ธธ ์๋ ์์ต๋๋ค. ์ด ๊ฒฝ์ฐ, ์ํ์ค๋ฅผ ๋ ์งง๊ฒ ์ค์ผ ํ์๊ฐ ์์ต๋๋ค.
๋ชจ๋ธ์์ ํ์ฉํ๋ ์ต๋ ๊ธธ์ด๋ก ์ํ์ค๋ฅผ ์๋ฅด๋ ค๋ฉด `truncation` ๋งค๊ฐ๋ณ์๋ฅผ `True`๋ก ์ค์ ํ์ธ์:
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch_sentences, padding=True, truncation=True)
>>> print(encoded_input)
{'input_ids': [[101, 1252, 1184, 1164, 1248, 6462, 136, 102, 0, 0, 0, 0, 0, 0, 0],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102, 0, 0, 0, 0, 0, 0, 0, 0]],
'token_type_ids': [[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]],
'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]]}
```
<Tip>
๋ค์ํ ํจ๋ฉ๊ณผ ์๋ผ๋ด๊ธฐ ์ธ์์ ๋ํด ๋ ์์๋ณด๋ ค๋ฉด [ํจ๋ฉ๊ณผ ์๋ผ๋ด๊ธฐ](./pad_truncation) ๊ฐ๋
๊ฐ์ด๋๋ฅผ ํ์ธํด๋ณด์ธ์.
</Tip>
### ํ
์ ๋ง๋ค๊ธฐ[[build-tensors]]
๋ง์ง๋ง์ผ๋ก, ํ ํฌ๋์ด์ ๊ฐ ๋ชจ๋ธ์ ๊ณต๊ธ๋๋ ์ค์ ํ
์๋ฅผ ๋ฐํํ๋๋ก ํฉ๋๋ค.
`return_tensors` ๋งค๊ฐ๋ณ์๋ฅผ PyTorch์ ๊ฒฝ์ฐ `pt`, TensorFlow์ ๊ฒฝ์ฐ `tf`๋ก ์ค์ ํ์ธ์:
<frameworkcontent>
<pt>
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch_sentences, padding=True, truncation=True, return_tensors="pt")
>>> print(encoded_input)
{'input_ids': tensor([[101, 1252, 1184, 1164, 1248, 6462, 136, 102, 0, 0, 0, 0, 0, 0, 0],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102, 0, 0, 0, 0, 0, 0, 0, 0]]),
'token_type_ids': tensor([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]),
'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]])}
```
</pt>
<tf>
```py
>>> batch_sentences = [
... "But what about second breakfast?",
... "Don't think he knows about second breakfast, Pip.",
... "What about elevensies?",
... ]
>>> encoded_input = tokenizer(batch_sentences, padding=True, truncation=True, return_tensors="tf")
>>> print(encoded_input)
{'input_ids': <tf.Tensor: shape=(2, 9), dtype=int32, numpy=
array([[101, 1252, 1184, 1164, 1248, 6462, 136, 102, 0, 0, 0, 0, 0, 0, 0],
[101, 1790, 112, 189, 1341, 1119, 3520, 1164, 1248, 6462, 117, 21902, 1643, 119, 102],
[101, 1327, 1164, 5450, 23434, 136, 102, 0, 0, 0, 0, 0, 0, 0, 0]],
dtype=int32)>,
'token_type_ids': <tf.Tensor: shape=(2, 9), dtype=int32, numpy=
array([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]], dtype=int32)>,
'attention_mask': <tf.Tensor: shape=(2, 9), dtype=int32, numpy=
array([[1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0],
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0]], dtype=int32)>}
```
</tf>
</frameworkcontent>
## ์ค๋์ค[[audio]]
์ค๋์ค ์์
์ ๋ชจ๋ธ์ ๋ง๋ ๋ฐ์ดํฐ ์ธํธ๋ฅผ ์ค๋นํ๊ธฐ ์ํด [ํน์ฑ ์ถ์ถ๊ธฐ](main_classes/feature_extractor)๊ฐ ํ์ํฉ๋๋ค. ํน์ฑ ์ถ์ถ๊ธฐ๋ ์์ ์ค๋์ค ๋ฐ์ดํฐ์์ ํน์ฑ๋ฅผ ์ถ์ถํ๊ณ ์ด๋ฅผ ํ
์๋ก ๋ณํํ๋ ๊ฒ์ด ๋ชฉ์ ์
๋๋ค.
์ค๋์ค ๋ฐ์ดํฐ ์ธํธ์ ํน์ฑ ์ถ์ถ๊ธฐ๋ฅผ ์ฌ์ฉํ๋ ๋ฐฉ๋ฒ์ ๋ณด๊ธฐ ์ํด [MInDS-14](https://huggingface.co/datasets/PolyAI/minds14) ๋ฐ์ดํฐ ์ธํธ๋ฅผ ๊ฐ์ ธ์ค์ธ์. (๋ฐ์ดํฐ ์ธํธ๋ฅผ ๊ฐ์ ธ์ค๋ ๋ฐฉ๋ฒ์ ๐ค [๋ฐ์ดํฐ ์ธํธ ํํ ๋ฆฌ์ผ](https://huggingface.co/docs/datasets/load_hub)์์ ์์ธํ ์ค๋ช
ํ๊ณ ์์ต๋๋ค.)
```py
>>> from datasets import load_dataset, Audio
>>> dataset = load_dataset("PolyAI/minds14", name="en-US", split="train")
```
`audio` ์ด์ ์ฒซ ๋ฒ์งธ ์์์ ์ ๊ทผํ์ฌ ์
๋ ฅ์ ์ดํด๋ณด์ธ์. `audio` ์ด์ ํธ์ถํ๋ฉด ์ค๋์ค ํ์ผ์ ์๋์ผ๋ก ๊ฐ์ ธ์ค๊ณ ๋ฆฌ์ํ๋งํฉ๋๋ค.
```py
>>> dataset[0]["audio"]
{'array': array([ 0. , 0.00024414, -0.00024414, ..., -0.00024414,
0. , 0. ], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/f14948e0e84be638dd7943ac36518a4cf3324e8b7aa331c5ab11541518e9368c/en-US~JOINT_ACCOUNT/602ba55abb1e6d0fbce92065.wav',
'sampling_rate': 8000}
```
์ด๋ ๊ฒ ํ๋ฉด ์ธ ๊ฐ์ง ํญ๋ชฉ์ด ๋ฐํ๋ฉ๋๋ค:
* `array`๋ 1D ๋ฐฐ์ด๋ก ๊ฐ์ ธ์์ (ํ์ํ ๊ฒฝ์ฐ) ๋ฆฌ์ํ๋ง๋ ์์ฑ ์ ํธ์
๋๋ค.
* `path`๋ ์ค๋์ค ํ์ผ์ ์์น๋ฅผ ๊ฐ๋ฆฌํต๋๋ค.
* `sampling_rate`๋ ์์ฑ ์ ํธ์์ ์ด๋น ์ธก์ ๋๋ ๋ฐ์ดํฐ ํฌ์ธํธ ์๋ฅผ ๋ํ๋
๋๋ค.
์ด ํํ ๋ฆฌ์ผ์์๋ [Wav2Vec2](https://huggingface.co/facebook/wav2vec2-base) ๋ชจ๋ธ์ ์ฌ์ฉํฉ๋๋ค. ๋ชจ๋ธ ์นด๋๋ฅผ ๋ณด๋ฉด Wav2Vec2๊ฐ 16kHz ์ํ๋ง๋ ์์ฑ ์ค๋์ค๋ฅผ ๊ธฐ๋ฐ์ผ๋ก ์ฌ์ ํ๋ จ๋ ๊ฒ์ ์ ์ ์์ต๋๋ค.
๋ชจ๋ธ์ ์ฌ์ ํ๋ จํ๋ ๋ฐ ์ฌ์ฉ๋ ๋ฐ์ดํฐ ์ธํธ์ ์ํ๋ง ๋ ์ดํธ์ ์ค๋์ค ๋ฐ์ดํฐ์ ์ํ๋ง ๋ ์ดํธ๊ฐ ์ผ์นํด์ผ ํฉ๋๋ค. ๋ฐ์ดํฐ์ ์ํ๋ง ๋ ์ดํธ๊ฐ ๋ค๋ฅด๋ฉด ๋ฐ์ดํฐ๋ฅผ ๋ฆฌ์ํ๋งํด์ผ ํฉ๋๋ค.
1. ๐ค Datasets์ [`~datasets.Dataset.cast_column`] ๋ฉ์๋๋ฅผ ์ฌ์ฉํ์ฌ ์ํ๋ง ๋ ์ดํธ๋ฅผ 16kHz๋ก ์
์ํ๋งํ์ธ์:
```py
>>> dataset = dataset.cast_column("audio", Audio(sampling_rate=16_000))
```
2. ์ค๋์ค ํ์ผ์ ๋ฆฌ์ํ๋งํ๊ธฐ ์ํด `audio` ์ด์ ๋ค์ ํธ์ถํฉ๋๋ค:
```py
>>> dataset[0]["audio"]
{'array': array([ 2.3443763e-05, 2.1729663e-04, 2.2145823e-04, ...,
3.8356509e-05, -7.3497440e-06, -2.1754686e-05], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/f14948e0e84be638dd7943ac36518a4cf3324e8b7aa331c5ab11541518e9368c/en-US~JOINT_ACCOUNT/602ba55abb1e6d0fbce92065.wav',
'sampling_rate': 16000}
```
๋ค์์ผ๋ก, ์
๋ ฅ์ ์ ๊ทํํ๊ณ ํจ๋ฉํ ํน์ฑ ์ถ์ถ๊ธฐ๋ฅผ ๊ฐ์ ธ์ค์ธ์. ํ
์คํธ ๋ฐ์ดํฐ์ ๊ฒฝ์ฐ, ๋ ์งง์ ์ํ์ค์ ๋ํด `0`์ด ์ถ๊ฐ๋ฉ๋๋ค. ์ค๋์ค ๋ฐ์ดํฐ์๋ ๊ฐ์ ๊ฐ๋
์ด ์ ์ฉ๋ฉ๋๋ค.
ํน์ฑ ์ถ์ถ๊ธฐ๋ ๋ฐฐ์ด์ `0`(๋ฌต์์ผ๋ก ํด์)์ ์ถ๊ฐํฉ๋๋ค.
[`AutoFeatureExtractor.from_pretrained`]๋ฅผ ์ฌ์ฉํ์ฌ ํน์ฑ ์ถ์ถ๊ธฐ๋ฅผ ๊ฐ์ ธ์ค์ธ์:
```py
>>> from transformers import AutoFeatureExtractor
>>> feature_extractor = AutoFeatureExtractor.from_pretrained("facebook/wav2vec2-base")
```
์ค๋์ค `array`๋ฅผ ํน์ฑ ์ถ์ถ๊ธฐ์ ์ ๋ฌํ์ธ์. ๋ํ, ๋ฐ์ํ ์ ์๋ ์กฐ์ฉํ ์ค๋ฅ(silent errors)๋ฅผ ๋ ์ ๋๋ฒ๊น
ํ ์ ์๋๋ก ํน์ฑ ์ถ์ถ๊ธฐ์ `sampling_rate` ์ธ์๋ฅผ ์ถ๊ฐํ๋ ๊ฒ์ ๊ถ์ฅํฉ๋๋ค.
```py
>>> audio_input = [dataset[0]["audio"]["array"]]
>>> feature_extractor(audio_input, sampling_rate=16000)
{'input_values': [array([ 3.8106556e-04, 2.7506407e-03, 2.8015103e-03, ...,
5.6335266e-04, 4.6588284e-06, -1.7142107e-04], dtype=float32)]}
```
ํ ํฌ๋์ด์ ์ ๋ง์ฐฌ๊ฐ์ง๋ก ๋ฐฐ์น ๋ด์์ ๊ฐ๋ณ์ ์ธ ์ํ์ค๋ฅผ ์ฒ๋ฆฌํ๊ธฐ ์ํด ํจ๋ฉ ๋๋ ์๋ผ๋ด๊ธฐ๋ฅผ ์ ์ฉํ ์ ์์ต๋๋ค. ์ด ๋ ๊ฐ์ ์ค๋์ค ์ํ์ ์ํ์ค ๊ธธ์ด๋ฅผ ํ์ธํด๋ณด์ธ์:
```py
>>> dataset[0]["audio"]["array"].shape
(173398,)
>>> dataset[1]["audio"]["array"].shape
(106496,)
```
์ค๋์ค ์ํ์ ๊ธธ์ด๊ฐ ๋์ผํ๋๋ก ๋ฐ์ดํฐ ์ธํธ๋ฅผ ์ ์ฒ๋ฆฌํ๋ ํจ์๋ฅผ ๋ง๋์ธ์. ์ต๋ ์ํ ๊ธธ์ด๋ฅผ ์ง์ ํ๋ฉด ํน์ฑ ์ถ์ถ๊ธฐ๊ฐ ํด๋น ๊ธธ์ด์ ๋ง์ถฐ ์ํ์ค๋ฅผ ํจ๋ฉํ๊ฑฐ๋ ์๋ผ๋
๋๋ค:
```py
>>> def preprocess_function(examples):
... audio_arrays = [x["array"] for x in examples["audio"]]
... inputs = feature_extractor(
... audio_arrays,
... sampling_rate=16000,
... padding=True,
... max_length=100000,
... truncation=True,
... )
... return inputs
```
`preprocess_function`์ ๋ฐ์ดํฐ ์ธํธ์ ์ฒ์ ์์ ๋ช ๊ฐ์ ์ ์ฉํด๋ณด์ธ์:
```py
>>> processed_dataset = preprocess_function(dataset[:5])
```
์ด์ ์ํ ๊ธธ์ด๊ฐ ๋ชจ๋ ๊ฐ๊ณ ์ง์ ๋ ์ต๋ ๊ธธ์ด์ ๋ง๊ฒ ๋์์ต๋๋ค. ๋๋์ด ์ ์ฒ๋ฆฌ๋ ๋ฐ์ดํฐ ์ธํธ๋ฅผ ๋ชจ๋ธ์ ์ ๋ฌํ ์ ์์ต๋๋ค!
```py
>>> processed_dataset["input_values"][0].shape
(100000,)
>>> processed_dataset["input_values"][1].shape
(100000,)
```
## ์ปดํจํฐ ๋น์ [[computer-vision]]
์ปดํจํฐ ๋น์ ์์
์ ๊ฒฝ์ฐ, ๋ชจ๋ธ์ ๋ํ ๋ฐ์ดํฐ ์ธํธ๋ฅผ ์ค๋นํ๊ธฐ ์ํด [์ด๋ฏธ์ง ํ๋ก์ธ์](main_classes/image_processor)๊ฐ ํ์ํฉ๋๋ค.
์ด๋ฏธ์ง ์ ์ฒ๋ฆฌ๋ ์ด๋ฏธ์ง๋ฅผ ๋ชจ๋ธ์ด ์์ํ๋ ์
๋ ฅ์ผ๋ก ๋ณํํ๋ ์ฌ๋ฌ ๋จ๊ณ๋ก ์ด๋ฃจ์ด์ง๋๋ค.
์ด๋ฌํ ๋จ๊ณ์๋ ํฌ๊ธฐ ์กฐ์ , ์ ๊ทํ, ์์ ์ฑ๋ ๋ณด์ , ์ด๋ฏธ์ง์ ํ
์ ๋ณํ ๋ฑ์ด ํฌํจ๋ฉ๋๋ค.
<Tip>
์ด๋ฏธ์ง ์ ์ฒ๋ฆฌ๋ ์ด๋ฏธ์ง ์ฆ๊ฐ ๊ธฐ๋ฒ์ ๋ช ๊ฐ์ง ์ ์ฉํ ๋ค์ ํ ์๋ ์์ต๋๋ค.
์ด๋ฏธ์ง ์ ์ฒ๋ฆฌ ๋ฐ ์ด๋ฏธ์ง ์ฆ๊ฐ์ ๋ชจ๋ ์ด๋ฏธ์ง ๋ฐ์ดํฐ๋ฅผ ๋ณํํ์ง๋ง, ์๋ก ๋ค๋ฅธ ๋ชฉ์ ์ ๊ฐ์ง๊ณ ์์ต๋๋ค:
* ์ด๋ฏธ์ง ์ฆ๊ฐ์ ๊ณผ์ ํฉ(over-fitting)์ ๋ฐฉ์งํ๊ณ ๋ชจ๋ธ์ ๊ฒฌ๊ณ ํจ(resiliency)์ ๋์ด๋ ๋ฐ ๋์์ด ๋๋ ๋ฐฉ์์ผ๋ก ์ด๋ฏธ์ง๋ฅผ ์์ ํฉ๋๋ค.
๋ฐ๊ธฐ์ ์์ ์กฐ์ , ์๋ฅด๊ธฐ, ํ์ , ํฌ๊ธฐ ์กฐ์ , ํ๋/์ถ์ ๋ฑ ๋ค์ํ ๋ฐฉ๋ฒ์ผ๋ก ๋ฐ์ดํฐ๋ฅผ ์ฆ๊ฐํ ์ ์์ต๋๋ค.
๊ทธ๋ฌ๋ ์ฆ๊ฐ์ผ๋ก ์ด๋ฏธ์ง์ ์๋ฏธ๊ฐ ๋ฐ๋์ง ์๋๋ก ์ฃผ์ํด์ผ ํฉ๋๋ค.
* ์ด๋ฏธ์ง ์ ์ฒ๋ฆฌ๋ ์ด๋ฏธ์ง๊ฐ ๋ชจ๋ธ์ด ์์ํ๋ ์
๋ ฅ ํ์๊ณผ ์ผ์นํ๋๋ก ๋ณด์ฅํฉ๋๋ค.
์ปดํจํฐ ๋น์ ๋ชจ๋ธ์ ๋ฏธ์ธ ์กฐ์ ํ ๋ ์ด๋ฏธ์ง๋ ๋ชจ๋ธ์ด ์ด๊ธฐ์ ํ๋ จ๋ ๋์ ์ ํํ ๊ฐ์ ๋ฐฉ์์ผ๋ก ์ ์ฒ๋ฆฌ๋์ด์ผ ํฉ๋๋ค.
์ด๋ฏธ์ง ์ฆ๊ฐ์๋ ์ํ๋ ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ๋ฌด์์ด๋ ์ฌ์ฉํ ์ ์์ต๋๋ค. ์ด๋ฏธ์ง ์ ์ฒ๋ฆฌ์๋ ๋ชจ๋ธ๊ณผ ์ฐ๊ฒฐ๋ `ImageProcessor`๋ฅผ ์ฌ์ฉํฉ๋๋ค.
</Tip>
[food101](https://huggingface.co/datasets/food101) ๋ฐ์ดํฐ ์ธํธ๋ฅผ ๊ฐ์ ธ์์ ์ปดํจํฐ ๋น์ ๋ฐ์ดํฐ ์ธํธ์์ ์ด๋ฏธ์ง ํ๋ก์ธ์๋ฅผ ์ด๋ป๊ฒ ์ฌ์ฉํ๋์ง ์์๋ณด์ธ์.
๋ฐ์ดํฐ ์ธํธ๋ฅผ ๋ถ๋ฌ์ค๋ ๋ฐฉ๋ฒ์ ๐ค [๋ฐ์ดํฐ ์ธํธ ํํ ๋ฆฌ์ผ](https://huggingface.co/docs/datasets/load_hub)์ ์ฐธ๊ณ ํ์ธ์.
<Tip>
๋ฐ์ดํฐ ์ธํธ๊ฐ ์๋นํ ํฌ๊ธฐ ๋๋ฌธ์ ๐ค Datasets์ `split` ๋งค๊ฐ๋ณ์๋ฅผ ์ฌ์ฉํ์ฌ ํ๋ จ ์ธํธ์์ ์์ ์ํ๋ง ๊ฐ์ ธ์ค์ธ์!
</Tip>
```py
>>> from datasets import load_dataset
>>> dataset = load_dataset("food101", split="train[:100]")
```
๋ค์์ผ๋ก, ๐ค Datasets์ [`image`](https://huggingface.co/docs/datasets/package_reference/main_classes?highlight=image#datasets.Image)๋ก ์ด๋ฏธ์ง๋ฅผ ํ์ธํด๋ณด์ธ์:
```py
>>> dataset[0]["image"]
```
<div class="flex justify-center">
<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/vision-preprocess-tutorial.png"/>
</div>
[`AutoImageProcessor.from_pretrained`]๋ก ์ด๋ฏธ์ง ํ๋ก์ธ์๋ฅผ ๊ฐ์ ธ์ค์ธ์:
```py
>>> from transformers import AutoImageProcessor
>>> image_processor = AutoImageProcessor.from_pretrained("google/vit-base-patch16-224")
```
๋จผ์ ์ด๋ฏธ์ง ์ฆ๊ฐ ๋จ๊ณ๋ฅผ ์ถ๊ฐํด ๋ด
์๋ค. ์๋ฌด ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ ์ฌ์ฉํด๋ ๊ด์ฐฎ์ง๋ง, ์ด๋ฒ ํํ ๋ฆฌ์ผ์์๋ torchvision์ [`transforms`](https://pytorch.org/vision/stable/transforms.html) ๋ชจ๋์ ์ฌ์ฉํ๊ฒ ์ต๋๋ค.
๋ค๋ฅธ ๋ฐ์ดํฐ ์ฆ๊ฐ ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ์ฌ์ฉํด๋ณด๊ณ ์ถ๋ค๋ฉด, [Albumentations](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/image_classification_albumentations.ipynb) ๋๋ [Kornia notebooks](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/image_classification_kornia.ipynb)์์ ์ด๋ป๊ฒ ์ฌ์ฉํ๋์ง ๋ฐฐ์ธ ์ ์์ต๋๋ค.
1. [`Compose`](https://pytorch.org/vision/master/generated/torchvision.transforms.Compose.html)๋ก [`RandomResizedCrop`](https://pytorch.org/vision/main/generated/torchvision.transforms.RandomResizedCrop.html)์ [`ColorJitter`](https://pytorch.org/vision/main/generated/torchvision.transforms.ColorJitter.html) ๋ฑ ๋ณํ์ ๋ช ๊ฐ์ง ์ฐ๊ฒฐํ์ธ์.
์ฐธ๊ณ ๋ก ํฌ๊ธฐ ์กฐ์ ์ ํ์ํ ์ด๋ฏธ์ง์ ํฌ๊ธฐ ์๊ตฌ์ฌํญ์ `image_processor`์์ ๊ฐ์ ธ์ฌ ์ ์์ต๋๋ค.
์ผ๋ถ ๋ชจ๋ธ์ ์ ํํ ๋์ด์ ๋๋น๋ฅผ ์๊ตฌํ์ง๋ง, ์ ์ผ ์งง์ ๋ณ์ ๊ธธ์ด(`shortest_edge`)๋ง ์ ์๋ ๋ชจ๋ธ๋ ์์ต๋๋ค.
```py
>>> from torchvision.transforms import RandomResizedCrop, ColorJitter, Compose
>>> size = (
... image_processor.size["shortest_edge"]
... if "shortest_edge" in image_processor.size
... else (image_processor.size["height"], image_processor.size["width"])
... )
>>> _transforms = Compose([RandomResizedCrop(size), ColorJitter(brightness=0.5, hue=0.5)])
```
2. ๋ชจ๋ธ์ ์
๋ ฅ์ผ๋ก [`pixel_values`](model_doc/visionencoderdecoder#transformers.VisionEncoderDecoderModel.forward.pixel_values)๋ฅผ ๋ฐ์ต๋๋ค.
`ImageProcessor`๋ ์ด๋ฏธ์ง ์ ๊ทํ ๋ฐ ์ ์ ํ ํ
์ ์์ฑ์ ์ฒ๋ฆฌํ ์ ์์ต๋๋ค.
๋ฐฐ์น ์ด๋ฏธ์ง์ ๋ํ ์ด๋ฏธ์ง ์ฆ๊ฐ ๋ฐ ์ด๋ฏธ์ง ์ ์ฒ๋ฆฌ๋ฅผ ๊ฒฐํฉํ๊ณ `pixel_values`๋ฅผ ์์ฑํ๋ ํจ์๋ฅผ ๋ง๋ญ๋๋ค:
```py
>>> def transforms(examples):
... images = [_transforms(img.convert("RGB")) for img in examples["image"]]
... examples["pixel_values"] = image_processor(images, do_resize=False, return_tensors="pt")["pixel_values"]
... return examples
```
<Tip>
์์ ์์์๋ ์ด๋ฏธ์ง ์ฆ๊ฐ ์ค์ ์ด๋ฏธ์ง ํฌ๊ธฐ๋ฅผ ์กฐ์ ํ๊ธฐ ๋๋ฌธ์ `do_resize=False`๋ก ์ค์ ํ๊ณ , ํด๋น `image_processor`์์ `size` ์์ฑ์ ํ์ฉํ์ต๋๋ค.
์ด๋ฏธ์ง ์ฆ๊ฐ ์ค์ ์ด๋ฏธ์ง ํฌ๊ธฐ๋ฅผ ์กฐ์ ํ์ง ์์ ๊ฒฝ์ฐ ์ด ๋งค๊ฐ๋ณ์๋ฅผ ์๋ตํ์ธ์.
๊ธฐ๋ณธ์ ์ผ๋ก๋ `ImageProcessor`๊ฐ ํฌ๊ธฐ ์กฐ์ ์ ์ฒ๋ฆฌํฉ๋๋ค.
์ฆ๊ฐ ๋ณํ ๊ณผ์ ์์ ์ด๋ฏธ์ง๋ฅผ ์ ๊ทํํ๋ ค๋ฉด `image_processor.image_mean` ๋ฐ `image_processor.image_std` ๊ฐ์ ์ฌ์ฉํ์ธ์.
</Tip>
3. ๐ค Datasets์ [`set_transform`](https://huggingface.co/docs/datasets/process#format-transform)๋ฅผ ์ฌ์ฉํ์ฌ ์ค์๊ฐ์ผ๋ก ๋ณํ์ ์ ์ฉํฉ๋๋ค:
```py
>>> dataset.set_transform(transforms)
```
4. ์ด์ ์ด๋ฏธ์ง์ ์ ๊ทผํ๋ฉด ์ด๋ฏธ์ง ํ๋ก์ธ์๊ฐ `pixel_values`๋ฅผ ์ถ๊ฐํ ๊ฒ์ ์ ์ ์์ต๋๋ค.
๋๋์ด ์ฒ๋ฆฌ๋ ๋ฐ์ดํฐ ์ธํธ๋ฅผ ๋ชจ๋ธ์ ์ ๋ฌํ ์ ์์ต๋๋ค!
```py
>>> dataset[0].keys()
```
๋ค์์ ๋ณํ์ด ์ ์ฉ๋ ํ์ ์ด๋ฏธ์ง์
๋๋ค. ์ด๋ฏธ์ง๊ฐ ๋ฌด์์๋ก ์๋ ค๋๊ฐ๊ณ ์์ ์์ฑ์ด ๋ค๋ฆ
๋๋ค.
```py
>>> import numpy as np
>>> import matplotlib.pyplot as plt
>>> img = dataset[0]["pixel_values"]
>>> plt.imshow(img.permute(1, 2, 0))
```
<div class="flex justify-center">
<img src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/preprocessed_image.png"/>
</div>
<Tip>
`ImageProcessor`๋ ๊ฐ์ฒด ๊ฐ์ง, ์๋งจํฑ ์ธ๊ทธ๋ฉํ
์ด์
(semantic segmentation), ์ธ์คํด์ค ์ธ๊ทธ๋ฉํ
์ด์
(instance segmentation), ํ๋ํฑ ์ธ๊ทธ๋ฉํ
์ด์
(panoptic segmentation)๊ณผ ๊ฐ์ ์์
์ ๋ํ ํ์ฒ๋ฆฌ ๋ฐฉ๋ฒ์ ์ ๊ณตํฉ๋๋ค.
์ด๋ฌํ ๋ฐฉ๋ฒ์ ๋ชจ๋ธ์ ์์ ์ถ๋ ฅ์ ๊ฒฝ๊ณ ์์๋ ์ธ๊ทธ๋ฉํ
์ด์
๋งต๊ณผ ๊ฐ์ ์๋ฏธ ์๋ ์์ธก์ผ๋ก ๋ณํํด์ค๋๋ค.
</Tip>
### ํจ๋ฉ[[pad]]
์๋ฅผ ๋ค์ด, [DETR](./model_doc/detr)์ ๊ฐ์ ๊ฒฝ์ฐ์๋ ๋ชจ๋ธ์ด ํ๋ จํ ๋ ํฌ๊ธฐ ์กฐ์ ์ฆ๊ฐ์ ์ ์ฉํฉ๋๋ค.
์ด๋ก ์ธํด ๋ฐฐ์น ๋ด ์ด๋ฏธ์ง ํฌ๊ธฐ๊ฐ ๋ฌ๋ผ์ง ์ ์์ต๋๋ค.
[`DetrImageProcessor`]์ [`DetrImageProcessor.pad`]๋ฅผ ์ฌ์ฉํ๊ณ ์ฌ์ฉ์ ์ ์ `collate_fn`์ ์ ์ํด์ ๋ฐฐ์น ์ด๋ฏธ์ง๋ฅผ ์ฒ๋ฆฌํ ์ ์์ต๋๋ค.
```py
>>> def collate_fn(batch):
... pixel_values = [item["pixel_values"] for item in batch]
... encoding = image_processor.pad(pixel_values, return_tensors="pt")
... labels = [item["labels"] for item in batch]
... batch = {}
... batch["pixel_values"] = encoding["pixel_values"]
... batch["pixel_mask"] = encoding["pixel_mask"]
... batch["labels"] = labels
... return batch
```
## ๋ฉํฐ๋ชจ๋ฌ[[multimodal]]
๋ฉํฐ๋ชจ๋ฌ ์
๋ ฅ์ด ํ์ํ ์์
์ ๊ฒฝ์ฐ, ๋ชจ๋ธ์ ๋ฐ์ดํฐ ์ธํธ๋ฅผ ์ค๋นํ๊ธฐ ์ํ [ํ๋ก์ธ์](main_classes/processors)๊ฐ ํ์ํฉ๋๋ค.
ํ๋ก์ธ์๋ ํ ํฌ๋์ด์ ์ ํน์ฑ ์ถ์ถ๊ธฐ์ ๊ฐ์ ๋ ๊ฐ์ง ์ฒ๋ฆฌ ๊ฐ์ฒด๋ฅผ ๊ฒฐํฉํฉ๋๋ค.
[LJ Speech](https://huggingface.co/datasets/lj_speech) ๋ฐ์ดํฐ ์ธํธ๋ฅผ ๊ฐ์ ธ์์ ์๋ ์์ฑ ์ธ์(ASR)์ ์ํ ํ๋ก์ธ์๋ฅผ ์ฌ์ฉํ๋ ๋ฐฉ๋ฒ์ ํ์ธํ์ธ์.
(๋ฐ์ดํฐ ์ธํธ๋ฅผ ๊ฐ์ ธ์ค๋ ๋ฐฉ๋ฒ์ ๋ํ ์์ธํ ๋ด์ฉ์ ๐ค [๋ฐ์ดํฐ ์ธํธ ํํ ๋ฆฌ์ผ](https://huggingface.co/docs/datasets/load_hub)์์ ๋ณผ ์ ์์ต๋๋ค.)
```py
>>> from datasets import load_dataset
>>> lj_speech = load_dataset("lj_speech", split="train")
```
์๋ ์์ฑ ์ธ์(ASR)์์๋ `audio`์ `text`์๋ง ์ง์คํ๋ฉด ๋๋ฏ๋ก, ๋ค๋ฅธ ์ด๋ค์ ์ ๊ฑฐํ ์ ์์ต๋๋ค:
```py
>>> lj_speech = lj_speech.map(remove_columns=["file", "id", "normalized_text"])
```
์ด์ `audio`์ `text`์ด์ ์ดํด๋ณด์ธ์:
```py
>>> lj_speech[0]["audio"]
{'array': array([-7.3242188e-04, -7.6293945e-04, -6.4086914e-04, ...,
7.3242188e-04, 2.1362305e-04, 6.1035156e-05], dtype=float32),
'path': '/root/.cache/huggingface/datasets/downloads/extracted/917ece08c95cf0c4115e45294e3cd0dee724a1165b7fc11798369308a465bd26/LJSpeech-1.1/wavs/LJ001-0001.wav',
'sampling_rate': 22050}
>>> lj_speech[0]["text"]
'Printing, in the only sense with which we are at present concerned, differs from most if not from all the arts and crafts represented in the Exhibition'
```
๊ธฐ์กด์ ์ฌ์ ํ๋ จ๋ ๋ชจ๋ธ์์ ์ฌ์ฉ๋ ๋ฐ์ดํฐ ์ธํธ์ ์๋ก์ด ์ค๋์ค ๋ฐ์ดํฐ ์ธํธ์ ์ํ๋ง ๋ ์ดํธ๋ฅผ ์ผ์น์ํค๊ธฐ ์ํด ์ค๋์ค ๋ฐ์ดํฐ ์ธํธ์ ์ํ๋ง ๋ ์ดํธ๋ฅผ [๋ฆฌ์ํ๋ง](preprocessing#audio)ํด์ผ ํฉ๋๋ค!
```py
>>> lj_speech = lj_speech.cast_column("audio", Audio(sampling_rate=16_000))
```
[`AutoProcessor.from_pretrained`]๋ก ํ๋ก์ธ์๋ฅผ ๊ฐ์ ธ์ค์ธ์:
```py
>>> from transformers import AutoProcessor
>>> processor = AutoProcessor.from_pretrained("facebook/wav2vec2-base-960h")
```
1. `array`์ ๋ค์ด ์๋ ์ค๋์ค ๋ฐ์ดํฐ๋ฅผ `input_values`๋ก ๋ณํํ๊ณ `text`๋ฅผ ํ ํฐํํ์ฌ `labels`๋ก ๋ณํํ๋ ํจ์๋ฅผ ๋ง๋ญ๋๋ค.
๋ชจ๋ธ์ ์
๋ ฅ์ ๋ค์๊ณผ ๊ฐ์ต๋๋ค:
```py
>>> def prepare_dataset(example):
... audio = example["audio"]
... example.update(processor(audio=audio["array"], text=example["text"], sampling_rate=16000))
... return example
```
2. ์ํ์ `prepare_dataset` ํจ์์ ์ ์ฉํ์ธ์:
```py
>>> prepare_dataset(lj_speech[0])
```
์ด์ ํ๋ก์ธ์๊ฐ `input_values`์ `labels`๋ฅผ ์ถ๊ฐํ๊ณ , ์ํ๋ง ๋ ์ดํธ๋ ์ฌ๋ฐ๋ฅด๊ฒ 16kHz๋ก ๋ค์ด์ํ๋งํ์ต๋๋ค.
๋๋์ด ์ฒ๋ฆฌ๋ ๋ฐ์ดํฐ ์ธํธ๋ฅผ ๋ชจ๋ธ์ ์ ๋ฌํ ์ ์์ต๋๋ค!
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/pad_truncation.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ํจ๋ฉ๊ณผ ์๋ผ๋ด๊ธฐ[[padding-and-truncation]]
๋ฐฐ์น ์
๋ ฅ์ ๊ธธ์ด๊ฐ ๋ค๋ฅธ ๊ฒฝ์ฐ๊ฐ ๋ง์์ ๊ณ ์ ํฌ๊ธฐ ํ
์๋ก ๋ณํํ ์ ์์ต๋๋ค. ํจ๋ฉ๊ณผ ์๋ผ๋ด๊ธฐ๋ ๋ค์ํ ๊ธธ์ด์ ๋ฐฐ์น์์ ์ง์ฌ๊ฐํ ํ
์๋ฅผ ์์ฑํ ์ ์๋๋ก ์ด ๋ฌธ์ ๋ฅผ ํด๊ฒฐํ๋ ์ ๋ต์
๋๋ค. ํจ๋ฉ์ ํน์ํ **ํจ๋ฉ ํ ํฐ**์ ์ถ๊ฐํ์ฌ ์งง์ ์ํ์ค๊ฐ ๋ฐฐ์น์์ ๊ฐ์ฅ ๊ธด ์ํ์ค ๋๋ ๋ชจ๋ธ์์ ํ์ฉํ๋ ์ต๋ ๊ธธ์ด์ ๋์ผํ ๊ธธ์ด๋ฅผ ๊ฐ๋๋ก ํฉ๋๋ค. ์๋ผ๋ด๊ธฐ๋ ๊ธด ์ํ์ค๋ฅผ ์๋ผ๋ด์ด ํจ๋ฉ๊ณผ ๋ค๋ฅธ ๋ฐฉ์์ผ๋ก ์ํ์ค์ ๊ธธ์ด๋ฅผ ๋์ผํ๊ฒ ํฉ๋๋ค.
๋๋ถ๋ถ์ ๊ฒฝ์ฐ ๋ฐฐ์น์ ๊ฐ์ฅ ๊ธด ์ํ์ค์ ๊ธธ์ด๋ก ํจ๋ฉํ๊ณ ๋ชจ๋ธ์ด ํ์ฉํ ์ ์๋ ์ต๋ ๊ธธ์ด๋ก ์๋ผ๋ด๋ ๊ฒ์ด ์ ์๋ํฉ๋๋ค. ๊ทธ๋ฌ๋ ํ์ํ๋ค๋ฉด API๊ฐ ์ง์ํ๋ ๋ ๋ง์ ์ ๋ต์ ์ฌ์ฉํ ์ ์์ต๋๋ค. ํ์ํ ์ธ์๋ `padding`, `truncation`, `max_length` ์ธ ๊ฐ์ง์
๋๋ค.
`padding` ์ธ์๋ ํจ๋ฉ์ ์ ์ดํฉ๋๋ค. ๋ถ๋ฆฌ์ธ ๋๋ ๋ฌธ์์ด์ผ ์ ์์ต๋๋ค:
- `True` ๋๋ `'longest'`: ๋ฐฐ์น์์ ๊ฐ์ฅ ๊ธด ์ํ์ค๋ก ํจ๋ฉํฉ๋๋ค(๋จ์ผ ์ํ์ค๋ง ์ ๊ณตํ๋ ๊ฒฝ์ฐ ํจ๋ฉ์ด ์ ์ฉ๋์ง ์์ต๋๋ค).
- `'max_length'`: `max_length` ์ธ์๊ฐ ์ง์ ํ ๊ธธ์ด๋ก ํจ๋ฉํ๊ฑฐ๋, `max_length`๊ฐ ์ ๊ณต๋์ง ์์ ๊ฒฝ์ฐ(`max_length=None`) ๋ชจ๋ธ์์ ํ์ฉ๋๋ ์ต๋ ๊ธธ์ด๋ก ํจ๋ฉํฉ๋๋ค. ๋จ์ผ ์ํ์ค๋ง ์ ๊ณตํ๋ ๊ฒฝ์ฐ์๋ ํจ๋ฉ์ด ์ ์ฉ๋ฉ๋๋ค.
- `False` ๋๋ `'do_not_pad'`: ํจ๋ฉ์ด ์ ์ฉ๋์ง ์์ต๋๋ค. ์ด๊ฒ์ด ๊ธฐ๋ณธ ๋์์
๋๋ค.
`truncation` ์ธ์๋ ์๋ผ๋ผ ๋ฐฉ๋ฒ์ ์ ํฉ๋๋ค. ๋ถ๋ฆฌ์ธ ๋๋ ๋ฌธ์์ด์ผ ์ ์์ต๋๋ค:
- `True` ๋๋ `longest_first`: `max_length` ์ธ์๊ฐ ์ง์ ํ ์ต๋ ๊ธธ์ด๋ก ์๋ผ๋ด๊ฑฐ๋,
`max_length`๊ฐ ์ ๊ณต๋์ง ์์ ๊ฒฝ์ฐ(`max_length=None`) ๋ชจ๋ธ์์ ํ์ฉ๋๋ ์ต๋ ๊ธธ์ด๋ก ์๋ผ๋
๋๋ค.
์ํ์ค ์์์ ๊ฐ์ฅ ๊ธด ์ํ์ค์ ํ ํฐ์ ์ ์ ํ ๊ธธ์ด์ ๋๋ฌํ ๋๊น์ง ํ๋์ฉ ์ ๊ฑฐํฉ๋๋ค.
- `'only_second'`: `max_length` ์ธ์๊ฐ ์ง์ ํ ์ต๋ ๊ธธ์ด๋ก ์๋ผ๋ด๊ฑฐ๋,
`max_length`๊ฐ ์ ๊ณต๋์ง ์์ ๊ฒฝ์ฐ(`max_length=None`) ๋ชจ๋ธ์์ ํ์ฉ๋๋ ์ต๋ ๊ธธ์ด๋ก ์๋ผ๋
๋๋ค.
์ํ์ค ์(๋๋ ์ํ์ค ์์ ๋ฐฐ์น)๊ฐ ์ ๊ณต๋ ๊ฒฝ์ฐ ์์ ๋ ๋ฒ์งธ ๋ฌธ์ฅ๋ง ์๋ผ๋
๋๋ค.
- `'only_first'`: `max_length` ์ธ์๊ฐ ์ง์ ํ ์ต๋ ๊ธธ์ด๋ก ์๋ผ๋ด๊ฑฐ๋,
`max_length`๊ฐ ์ ๊ณต๋์ง ์์ ๊ฒฝ์ฐ(`max_length=None`) ๋ชจ๋ธ์์ ํ์ฉ๋๋ ์ต๋ ๊ธธ์ด๋ก ์๋ผ๋
๋๋ค.
์ํ์ค ์(๋๋ ์ํ์ค ์์ ๋ฐฐ์น)๊ฐ ์ ๊ณต๋ ๊ฒฝ์ฐ ์์ ์ฒซ ๋ฒ์งธ ๋ฌธ์ฅ๋ง ์๋ผ๋
๋๋ค.
- `False` ๋๋ `'do_not_truncate'`: ์๋ผ๋ด๊ธฐ๋ฅผ ์ ์ฉํ์ง ์์ต๋๋ค. ์ด๊ฒ์ด ๊ธฐ๋ณธ ๋์์
๋๋ค.
`max_length` ์ธ์๋ ํจ๋ฉ ๋ฐ ์๋ผ๋ด๊ธฐ๋ฅผ ์ ์ฉํ ๊ธธ์ด๋ฅผ ์ ์ดํฉ๋๋ค. ์ด ์ธ์๋ ์ ์ ๋๋ `None`์ผ ์ ์์ผ๋ฉฐ, `None`์ผ ๊ฒฝ์ฐ ๋ชจ๋ธ์ด ํ์ฉํ ์ ์๋ ์ต๋ ๊ธธ์ด๋ก ๊ธฐ๋ณธ๊ฐ์ด ์ค์ ๋ฉ๋๋ค. ๋ชจ๋ธ์ ํน์ ํ ์ต๋ ์
๋ ฅ ๊ธธ์ด๊ฐ ์๋ ๊ฒฝ์ฐ `max_length`์ ๋ํ ์๋ผ๋ด๊ธฐ ๋๋ ํจ๋ฉ์ด ๋นํ์ฑํ๋ฉ๋๋ค.
๋ค์ ํ์๋ ํจ๋ฉ ๋ฐ ์๋ผ๋ด๊ธฐ๋ฅผ ์ค์ ํ๋ ๊ถ์ฅ ๋ฐฉ๋ฒ์ด ์์ฝ๋์ด ์์ต๋๋ค.
์
๋ ฅ์ผ๋ก ์ํ์ค ์์ ์ฌ์ฉํ๋ ๊ฒฝ์ฐ, ๋ค์ ์์ ์์ `truncation=True`๋ฅผ `['only_first', 'only_second', 'longest_first']`์์ ์ ํํ `STRATEGY`, ์ฆ `truncation='only_second'` ๋๋ `truncation='longest_first'`๋ก ๋ฐ๊พธ๋ฉด ์์ ์ค๋ช
ํ ๋๋ก ์์ ๋ ์ํ์ค๊ฐ ์๋ฆฌ๋ ๋ฐฉ์์ ์ ์ดํ ์ ์์ต๋๋ค.
| ์๋ผ๋ด๊ธฐ | ํจ๋ฉ | ์ฌ์ฉ ๋ฐฉ๋ฒ |
|--------------------------------------|-----------------------------------|------------------------------------------------------------------------------------------|
| ์๋ผ๋ด๊ธฐ ์์ | ํจ๋ฉ ์์ | `tokenizer(batch_sentences)` |
| | ๋ฐฐ์น ๋ด ์ต๋ ๊ธธ์ด๋ก ํจ๋ฉ | `tokenizer(batch_sentences, padding=True)` ๋๋ |
| | | `tokenizer(batch_sentences, padding='longest')` |
| | ๋ชจ๋ธ์ ์ต๋ ์
๋ ฅ ๊ธธ์ด๋ก ํจ๋ฉ | `tokenizer(batch_sentences, padding='max_length')` |
| | ํน์ ๊ธธ์ด๋ก ํจ๋ฉ | `tokenizer(batch_sentences, padding='max_length', max_length=42)` |
| | ๋ค์ํ ๊ธธ์ด๋ก ํจ๋ฉ | `tokenizer(batch_sentences, padding=True, pad_to_multiple_of=8) |
| ๋ชจ๋ธ์ ์ต๋ ์
๋ ฅ ๊ธธ์ด๋ก ์๋ผ๋ด๊ธฐ | ํจ๋ฉ ์์ | `tokenizer(batch_sentences, truncation=True)` ๋๋ |
| | | `tokenizer(batch_sentences, truncation=STRATEGY)` |
| | ๋ฐฐ์น ๋ด ์ต๋ ๊ธธ์ด๋ก ํจ๋ฉ | `tokenizer(batch_sentences, padding=True, truncation=True)` ๋๋ |
| | | `tokenizer(batch_sentences, padding=True, truncation=STRATEGY)` |
| | ๋ชจ๋ธ์ ์ต๋ ์
๋ ฅ ๊ธธ์ด๋ก ํจ๋ฉ | `tokenizer(batch_sentences, padding='max_length', truncation=True)` ๋๋ |
| | | `tokenizer(batch_sentences, padding='max_length', truncation=STRATEGY)` |
| | ํน์ ๊ธธ์ด๋ก ํจ๋ฉ | ์ฌ์ฉ ๋ถ๊ฐ |
| ํน์ ๊ธธ์ด๋ก ์๋ผ๋ด๊ธฐ | ํจ๋ฉ ์์ | `tokenizer(batch_sentences, truncation=True, max_length=42)` ๋๋ |
| | | `tokenizer(batch_sentences, truncation=STRATEGY, max_length=42)` |
| | ๋ฐฐ์น ๋ด ์ต๋ ๊ธธ์ด๋ก ํจ๋ฉ | `tokenizer(batch_sentences, padding=True, truncation=True, max_length=42)` ๋๋ |
| | | `tokenizer(batch_sentences, padding=True, truncation=STRATEGY, max_length=42)` |
| | ๋ชจ๋ธ์ ์ต๋ ์
๋ ฅ ๊ธธ์ด๋ก ํจ๋ฉ | ์ฌ์ฉ ๋ถ๊ฐ |
| | ํน์ ๊ธธ์ด๋ก ํจ๋ฉ | `tokenizer(batch_sentences, padding='max_length', truncation=True, max_length=42)` ๋๋ |
| | | `tokenizer(batch_sentences, padding='max_length', truncation=STRATEGY, max_length=42)` |
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/in_translation.md
|
<!--โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ์ด์ฌํ ๋ฒ์ญ ์ค์
๋๋ค. ์กฐ๊ธ ์ด๋ฐ ๋ง๋์!
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/model_memory_anatomy.md
|
<!---
Copyright 2023 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->
# ๋ชจ๋ธ ํ์ต ํด๋ถํ๊ธฐ [[model-training-anatomy]]
๋ชจ๋ธ ํ๋ จ ์๋์ ๋ฉ๋ชจ๋ฆฌ ํ์ฉ์ ํจ์จ์ฑ์ ํฅ์์ํค๊ธฐ ์ํด ์ ์ฉํ ์ ์๋ ์ฑ๋ฅ ์ต์ ํ ๊ธฐ์ ์ ์ดํดํ๋ ค๋ฉด GPU๊ฐ ํ๋ จ ์ค์ ์ด๋ป๊ฒ ํ์ฉ๋๋์ง, ๊ทธ๋ฆฌ๊ณ ์ํ๋๋ ์ฐ์ฐ์ ๋ฐ๋ผ ์ฐ์ฐ ๊ฐ๋๊ฐ ์ด๋ป๊ฒ ๋ณํ๋์ง์ ์ต์ํด์ ธ์ผ ํฉ๋๋ค.
๋จผ์ GPU ํ์ฉ๊ณผ ๋ชจ๋ธ ํ๋ จ ์คํ์ ๋ํ ์์๋ฅผ ์ดํด๋ณด๊ฒ ์ต๋๋ค. ๋ฐ๋ชจ๋ฅผ ์ํด ๋ช๋ช ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ์ค์นํด์ผ ํฉ๋๋ค:
```bash
pip install transformers datasets accelerate nvidia-ml-py3
```
`nvidia-ml-py3` ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ Python ๋ด๋ถ์์ ๋ชจ๋ธ์ ๋ฉ๋ชจ๋ฆฌ ์ฌ์ฉ๋์ ๋ชจ๋ํฐ๋งํ ์ ์๊ฒ ํด์ค๋๋ค. ํฐ๋ฏธ๋์ `nvidia-smi` ๋ช
๋ น์ด์ ์ต์ํ ์ ์๋๋ฐ, ์ด ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ Python์์ ์ง์ ๋์ผํ ์ ๋ณด์ ์ ๊ทผํ ์ ์๊ฒ ํด์ค๋๋ค.
๊ทธ ๋ค์, 100๊ณผ 30000 ์ฌ์ด์ ๋ฌด์์ ํ ํฐ ID์ ๋ถ๋ฅ๊ธฐ๋ฅผ ์ํ ์ด์ง ๋ ์ด๋ธ์ธ ๋๋ฏธ ๋ฐ์ดํฐ๋ฅผ ์์ฑํฉ๋๋ค.
๊ธธ์ด๊ฐ ๊ฐ๊ฐ 512์ธ ์ด 512๊ฐ์ ์ํ์ค๋ฅผ ๊ฐ์ ธ์ PyTorch ํ์์ [`~datasets.Dataset`]์ ์ ์ฅํฉ๋๋ค.
```py
>>> import numpy as np
>>> from datasets import Dataset
>>> seq_len, dataset_size = 512, 512
>>> dummy_data = {
... "input_ids": np.random.randint(100, 30000, (dataset_size, seq_len)),
... "labels": np.random.randint(0, 1, (dataset_size)),
... }
>>> ds = Dataset.from_dict(dummy_data)
>>> ds.set_format("pt")
```
GPU ํ์ฉ ๋ฐ [`Trainer`]๋ก ์คํํ ํ๋ จ ๊ณผ์ ์ ๋ํ ์์ฝ ํต๊ณ๋ฅผ ์ถ๋ ฅํ๊ธฐ ์ํด ๋ ๊ฐ์ ๋์ฐ๋ฏธ ํจ์๋ฅผ ์ ์ํ๊ฒ ์ต๋๋ค:
```py
>>> from pynvml import *
>>> def print_gpu_utilization():
... nvmlInit()
... handle = nvmlDeviceGetHandleByIndex(0)
... info = nvmlDeviceGetMemoryInfo(handle)
... print(f"GPU memory occupied: {info.used//1024**2} MB.")
>>> def print_summary(result):
... print(f"Time: {result.metrics['train_runtime']:.2f}")
... print(f"Samples/second: {result.metrics['train_samples_per_second']:.2f}")
... print_gpu_utilization()
```
์์ํ ๋ GPU ๋ฉ๋ชจ๋ฆฌ๊ฐ ๋น์ด ์๋์ง ํ์ธํด ๋ด
์๋ค:
```py
>>> print_gpu_utilization()
GPU memory occupied: 0 MB.
```
์ข์ต๋๋ค. ๋ชจ๋ธ์ ๋ก๋ํ๊ธฐ ์ ์๋ ์์๋๋ก GPU ๋ฉ๋ชจ๋ฆฌ๊ฐ ์ ์ ๋์ง ์์์ต๋๋ค. ๊ทธ๋ ์ง ์๋ค๋ฉด ์ฌ์ฉ์์ ๊ธฐ๊ธฐ์์ GPU ๋ฉ๋ชจ๋ฆฌ๋ฅผ ์ฌ์ฉํ๋ ๋ชจ๋ ํ๋ก์ธ์ค๋ฅผ ์ค๋จํด์ผ ํฉ๋๋ค. ๊ทธ๋ฌ๋ ์ฌ์ฉ์๋ ๋ชจ๋ ์ฌ์ GPU ๋ฉ๋ชจ๋ฆฌ๋ฅผ ์ฌ์ฉํ ์๋ ์์ต๋๋ค. ๋ชจ๋ธ์ด GPU์ ๋ก๋๋ ๋ ์ปค๋๋ ๋ก๋๋๋ฏ๋ก 1-2GB์ ๋ฉ๋ชจ๋ฆฌ๋ฅผ ์ฐจ์งํ ์ ์์ต๋๋ค. ์ผ๋ง๋ ๋๋์ง ํ์ธํ๊ธฐ ์ํด GPU์ ์์ ํ
์๋ฅผ ๋ก๋ํ์ฌ ์ปค๋์ด ๋ก๋๋๋๋ก ํธ๋ฆฌ๊ฑฐํฉ๋๋ค.
```py
>>> import torch
>>> torch.ones((1, 1)).to("cuda")
>>> print_gpu_utilization()
GPU memory occupied: 1343 MB.
```
์ปค๋๋ง์ผ๋ก๋ GPU ๋ฉ๋ชจ๋ฆฌ์ 1.3GB๋ฅผ ์ฐจ์งํฉ๋๋ค. ์ด์ ๋ชจ๋ธ์ด ์ผ๋ง๋ ๋ง์ ๊ณต๊ฐ์ ์ฌ์ฉํ๋์ง ํ์ธํด ๋ณด๊ฒ ์ต๋๋ค.
## ๋ชจ๋ธ ๋ก๋ [[load-model]]
์ฐ์ , `bert-large-uncased` ๋ชจ๋ธ์ ๋ก๋ํฉ๋๋ค. ๋ชจ๋ธ์ ๊ฐ์ค์น๋ฅผ ์ง์ GPU์ ๋ก๋ํด์ ๊ฐ์ค์น๋ง์ด ์ผ๋ง๋ ๋ง์ ๊ณต๊ฐ์ ์ฐจ์งํ๋์ง ํ์ธํ ์ ์์ต๋๋ค.
```py
>>> from transformers import AutoModelForSequenceClassification
>>> model = AutoModelForSequenceClassification.from_pretrained("bert-large-uncased").to("cuda")
>>> print_gpu_utilization()
GPU memory occupied: 2631 MB.
```
๋ชจ๋ธ์ ๊ฐ์ค์น๋ง์ผ๋ก๋ GPU ๋ฉ๋ชจ๋ฆฌ๋ฅผ 1.3 GB ์ฐจ์งํ๋ ๊ฒ์ ๋ณผ ์ ์์ต๋๋ค. ์ ํํ ์ซ์๋ ์ฌ์ฉํ๋ GPU์ ๋ฐ๋ผ ๋ค๋ฆ
๋๋ค. ์ต์ GPU์์๋ ๋ชจ๋ธ ์ฌ์ฉ ์๋๋ฅผ ๋์ด๋ ์ต์ ํ๋ ๋ฐฉ์์ผ๋ก ๊ฐ์ค์น๊ฐ ๋ก๋๋๋ฏ๋ก, ๋ชจ๋ธ์ด ๋ ๋ง์ ๊ณต๊ฐ์ ์ฐจ์งํ ์ ์์ต๋๋ค. ์ด์ `nvidia-smi` CLI์ ๋์ผํ ๊ฒฐ๊ณผ๋ฅผ ์ป๋์ง ๋น ๋ฅด๊ฒ ํ์ธํ ์ ์์ต๋๋ค:
```bash
nvidia-smi
```
```bash
Tue Jan 11 08:58:05 2022
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.91.03 Driver Version: 460.91.03 CUDA Version: 11.2 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 Tesla V100-SXM2... On | 00000000:00:04.0 Off | 0 |
| N/A 37C P0 39W / 300W | 2631MiB / 16160MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| 0 N/A N/A 3721 C ...nvs/codeparrot/bin/python 2629MiB |
+-----------------------------------------------------------------------------+
```
์ด์ ๊ณผ ๋์ผํ ์ซ์๊ฐ ์ถ๋ ฅ๋๊ณ 16GB ๋ฉ๋ชจ๋ฆฌ๋ฅผ ๊ฐ์ง V100 GPU๋ฅผ ์ฌ์ฉํ๊ณ ์๋ค๋ ๊ฒ๋ ๋ณผ ์ ์์ต๋๋ค. ๊ทธ๋ฌ๋ฏ๋ก ์ด์ ๋ชจ๋ธ ํ๋ จ์ ์์ํ์ฌ GPU ๋ฉ๋ชจ๋ฆฌ ์ฌ์ฉ๋์ด ์ด๋ป๊ฒ ๋ฌ๋ผ์ง๋์ง ๋ณผ ์ ์์ต๋๋ค. ์ฐ์ ๋ช๋ช ํ์ค ํ๋ จ ์ธ์๋ฅผ ์ค์ ํฉ๋๋ค:
```py
default_args = {
"output_dir": "tmp",
"evaluation_strategy": "steps",
"num_train_epochs": 1,
"log_level": "error",
"report_to": "none",
}
```
<Tip>
์ฌ๋ฌ ์คํ์ ์คํํ ๊ณํ์ด๋ผ๋ฉด, ์คํ ๊ฐ์ ๋ฉ๋ชจ๋ฆฌ๋ฅผ ์ ๋๋ก ๋น์ฐ๊ธฐ ์ํด์ Python ์ปค๋์ ์คํ ์ฌ์ด๋ง๋ค ์ฌ์์ํด์ผ ํฉ๋๋ค.
</Tip>
## ๊ธฐ๋ณธ ํ๋ จ์์์ ๋ฉ๋ชจ๋ฆฌ ํ์ฉ [[memory-utilization-at-vanilla-training]]
[`Trainer`]๋ฅผ ์ฌ์ฉํ์ฌ, GPU ์ฑ๋ฅ ์ต์ ํ ๊ธฐ์ ์ ์ฌ์ฉํ์ง ์๊ณ ๋ฐฐ์น ํฌ๊ธฐ๊ฐ 4์ธ ๋ชจ๋ธ์ ํ๋ จ์ํค๊ฒ ์ต๋๋ค:
```py
>>> from transformers import TrainingArguments, Trainer, logging
>>> logging.set_verbosity_error()
>>> training_args = TrainingArguments(per_device_train_batch_size=4, **default_args)
>>> trainer = Trainer(model=model, args=training_args, train_dataset=ds)
>>> result = trainer.train()
>>> print_summary(result)
```
```
Time: 57.82
Samples/second: 8.86
GPU memory occupied: 14949 MB.
```
์ฐ๋ฆฌ๋ ๋น๊ต์ ์์ ๋ฐฐ์น ํฌ๊ธฐ๋ก๋ ์ ์ฒด GPU ๋ฉ๋ชจ๋ฆฌ๋ฅผ ๊ฑฐ์ ๋ค ์ฐจ์งํ๋ ๊ฒ์ ๋ณผ ์ ์์ต๋๋ค. ๊ทธ๋ฌ๋ ๋ฐฐ์น ํฌ๊ธฐ๊ฐ ํด์๋ก ๋ชจ๋ธ ์๋ ด ์๋๊ฐ ๋นจ๋ผ์ง๊ณ ์ต์ข
์ฑ๋ฅ์ด ํฅ์๋๋ ๊ฒฝ์ฐ๊ฐ ๋ง์ต๋๋ค. ๊ทธ๋์ ์ด์์ ์ผ๋ก๋ GPU ์ ํ์ด ์๋ ์ฐ๋ฆฌ ๋ชจ๋ธ์ ์๊ตฌ์ฌํญ์ ๋ง๊ฒ ๋ฐฐ์น ํฌ๊ธฐ๋ฅผ ์กฐ์ ํ๋ ค๊ณ ํฉ๋๋ค. ํฅ๋ฏธ๋กญ๊ฒ๋ ์ฐ๋ฆฌ๋ ๋ชจ๋ธ์ ํฌ๊ธฐ๋ณด๋ค ํจ์ฌ ๋ ๋ง์ ๋ฉ๋ชจ๋ฆฌ๋ฅผ ์ฌ์ฉํฉ๋๋ค. ์ ์ด๋ฐ ํ์์ด ๋ฐ์ํ๋์ง ์กฐ๊ธ ๋ ์ ์ดํดํ๊ธฐ ์ํด ๋ชจ๋ธ์ ์ฐ์ฐ๊ณผ ๋ฉ๋ชจ๋ฆฌ ์๊ตฌ ์ฌํญ์ ์ดํด๋ณด๊ฒ ์ต๋๋ค.
## ๋ชจ๋ธ์ ์ฐ์ฐ ํด๋ถํ๊ธฐ [[anatomy-of-models-operations]]
ํธ๋์คํฌ๋จธ ์ํคํ
์ฒ์๋ ์ฐ์ฐ ๊ฐ๋(compute-intensity)์ ๋ฐ๋ผ ๊ทธ๋ฃนํ๋ 3๊ฐ์ง ์ฃผ์ ์ฐ์ฐ ๊ทธ๋ฃน์ด ์์ต๋๋ค.
1. **ํ
์ ์ถ์ฝ(Tensor Contractions)**
์ ํ ๋ ์ด์ด์ ๋ฉํฐํค๋ ์ดํ
์
์ ๊ตฌ์ฑ ์์๋ ๋ชจ๋ **ํ๋ ฌ-ํ๋ ฌ ๊ณฑ์
(matrix-matrix multiplications)**์ ์ผ๊ด์ ์ผ๋ก ์ฒ๋ฆฌํฉ๋๋ค. ์ด ์ฐ์ฐ์ ํธ๋์คํฌ๋จธ ํ๋ จ์์ ๊ฐ์ฅ ์ฐ์ฐ ๊ฐ๋๊ฐ ๋์ ๋ถ๋ถ์
๋๋ค.
2. **ํต๊ณ ์ ๊ทํ(Statistical Normalizations)**
์ํํธ๋งฅ์ค์ ๋ ์ด์ด ์ ๊ทํ๋ ํ
์ ์ถ์ฝ๋ณด๋ค ์ฐ์ฐ ๊ฐ๋๊ฐ ๋ฎ์ต๋๋ค. ํ๋ ์ด์์ **๊ฐ์ ์ฐ์ฐ(reduction operations)**์ ํฌํจํ๋ฉฐ, ๊ทธ ๊ฒฐ๊ณผ๋ map์ ํตํด ์ ์ฉ๋ฉ๋๋ค.
3. **์์๋ณ ์ฐ์ฐ์(Element-wise Operators)**
๊ทธ ์ธ ์ฐ์ฐ์๋ค, **ํธํฅ(biases), ๋๋กญ์์(dropout), ํ์ฑํ ํจ์(activations), ์์ฐจ ์ฐ๊ฒฐ(residual connections)**์ด ์ฌ๊ธฐ์ ํด๋นํฉ๋๋ค. ์ด ์ฐ์ฐ๋ค์ ์ฐ์ฐ ๊ฐ๋๊ฐ ๊ฐ์ฅ ๋ฎ์ต๋๋ค.
์ด๋ฌํ ์ง์์ ์ฑ๋ฅ ๋ณ๋ชฉ ํ์์ ๋ถ์ํ ๋ ๋์์ด ๋ ์ ์์ต๋๋ค.
์ด ๋ด์ฉ์ [Data Movement Is All You Need: A Case Study on Optimizing Transformers 2020](https://arxiv.org/abs/2007.00072)์ ์ฐธ๊ณ ํ์์ต๋๋ค.
## ๋ชจ๋ธ์ ๋ฉ๋ชจ๋ฆฌ ๊ตฌ์กฐ [[anatomy-of-models-memory]]
๋ชจ๋ธ์ ํ๋ จ์ํค๋ ๋ฐ๋ ๋จ์ํ GPU์ ๋ชจ๋ธ์ ์ฌ๋ฆฌ๋ ๊ฒ๋ณด๋ค ํจ์ฌ ๋ ๋ง์ ๋ฉ๋ชจ๋ฆฌ๋ฅผ ์ฌ์ฉํ๋ค๋ ๊ฒ์ ๋ณด์์ต๋๋ค. ์ด๋ ํ๋ จ ์ค GPU ๋ฉ๋ชจ๋ฆฌ๋ฅผ ์ฌ์ฉํ๋ ๋ง์ ๊ตฌ์ฑ ์์๊ฐ ์๊ธฐ ๋๋ฌธ์
๋๋ค. GPU ๋ฉ๋ชจ๋ฆฌ์ ๊ตฌ์ฑ ์์๋ ๋ค์๊ณผ ๊ฐ์ต๋๋ค:
1. ๋ชจ๋ธ ๊ฐ์ค์น
2. ์ตํฐ๋ง์ด์ ์ํ
3. ๊ทธ๋ผ๋์ธํธ
4. ๊ทธ๋ผ๋์ธํธ ๊ณ์ฐ์ ์ํด ์ ์ฅ๋ ์๋ฐฉํฅ ํ์ฑํ
5. ์์ ๋ฒํผ
6. ๊ธฐ๋ฅ๋ณ ๋ฉ๋ชจ๋ฆฌ
AdamW๋ฅผ ์ฌ์ฉํ์ฌ ํผํฉ ์ ๋ฐ๋๋ก ํ๋ จ๋ ์ผ๋ฐ์ ์ธ ๋ชจ๋ธ์ ๋ชจ๋ธ ํ๋ผ๋ฏธํฐ๋น 18 ๋ฐ์ดํธ์ ํ์ฑํ ๋ฉ๋ชจ๋ฆฌ๊ฐ ํ์ํฉ๋๋ค. ์ถ๋ก ๋จ๊ณ์์๋ ์ตํฐ๋ง์ด์ ์ ๊ทธ๋ผ๋์ธํธ๊ฐ ํ์ํ์ง ์์ผ๋ฏ๋ก ์ด๋ค์ ์ ์ธํฉ๋๋ค. ๋ฐ๋ผ์ ํผํฉ ์ ๋ฐ๋ ์ถ๋ก ์ ๊ฒฝ์ฐ ๋ชจ๋ธ ๋งค๊ฐ๋ณ์๋น 6 ๋ฐ์ดํธ์ ํ์ฑํ ๋ฉ๋ชจ๋ฆฌ๊ฐ ํ์ํฉ๋๋ค.
์์ธํ ์ดํด๋ณด๊ฒ ์ต๋๋ค.
**๋ชจ๋ธ ๊ฐ์ค์น:**
- fp32 ํ๋ จ์ ๊ฒฝ์ฐ ๋งค๊ฐ ๋ณ์ ์ * 4 ๋ฐ์ดํธ
- ํผํฉ ์ ๋ฐ๋ ํ๋ จ์ ๊ฒฝ์ฐ ๋งค๊ฐ ๋ณ์ ์ * 6 ๋ฐ์ดํธ (๋ฉ๋ชจ๋ฆฌ์ fp32์ fp16 ๋ ๊ฐ์ง ๋ชจ๋ธ์ ์ ์ง)
**์ตํฐ๋ง์ด์ ์ํ:**
- ์ผ๋ฐ AdamW์ ๊ฒฝ์ฐ ๋งค๊ฐ ๋ณ์ ์ * 8 ๋ฐ์ดํธ (2๊ฐ์ง ์ํ ์ ์ง)
- [bitsandbytes](https://github.com/TimDettmers/bitsandbytes)์ ๊ฐ์ 8๋นํธ AdamW ์ตํฐ๋ง์ด์ ์ ๊ฒฝ์ฐ ๋งค๊ฐ ๋ณ์ ์ * 2 ๋ฐ์ดํธ
- Momentum์ ๊ฐ์ง SGD์ ๊ฐ์ ์ตํฐ๋ง์ด์ ์ ๊ฒฝ์ฐ ๋งค๊ฐ ๋ณ์ ์ * 4 ๋ฐ์ดํธ (ํ๋์ ์ํ๋ง ์ ์ง)
**๊ทธ๋ผ๋์ธํธ**
- fp32 ๋๋ ํผํฉ ์ ๋ฐ๋ ํ๋ จ์ ๊ฒฝ์ฐ ๋งค๊ฐ ๋ณ์ ์ * 4 ๋ฐ์ดํธ (๊ทธ๋ผ๋์ธํธ๋ ํญ์ fp32์ผ๋ก ์ ์ง๋ฉ๋๋ค.)
**์๋ฐฉํฅ ํ์ฑํ**
- ํฌ๊ธฐ๋ ์ฌ๋ฌ ์์ธ์ ๋ฐ๋ผ ๋ฌ๋ผ์ง๋ฉฐ, ์ฃผ์ ์์ธ์ ์ํ์ค ๊ธธ์ด, ์๋ ์ํ์ ํฌ๊ธฐ ๋ฐ ๋ฐฐ์น ํฌ๊ธฐ์
๋๋ค.
์๋ฐฉํฅ ๋ฐ ์ญ๋ฐฉํฅ ํจ์์์ ์ ๋ฌ ๋ฐ ๋ฐํ๋๋ ์
๋ ฅ๊ณผ ์ถ๋ ฅ์ด ์์ผ๋ฉฐ, ๊ทธ๋ผ๋์ธํธ ๊ณ์ฐ์ ์ํด ์ ์ฅ๋ ์๋ฐฉํฅ ํ์ฑํ๊ฐ ์์ต๋๋ค.
**์์ ๋ฉ๋ชจ๋ฆฌ**
๋๋ถ์ด ๋ชจ๋ ์ข
๋ฅ์ ์์ ๋ณ์๋ ์ฐ์ฐ์ด ์๋ฃ๋๋ฉด ๊ณง๋ฐ๋ก ํด์ ๋์ง๋ง, ๊ทธ ์๊ฐ์๋ ์ถ๊ฐ ๋ฉ๋ชจ๋ฆฌ๊ฐ ํ์ํ ์ ์๊ณ OOM์ ์ ๋ฐํ ์ ์์ต๋๋ค. ๋ฐ๋ผ์ ์ฝ๋ฉํ ๋ ์ด๋ฌํ ์์ ๋ณ์์ ๋ํด ์ ๋ต์ ์ผ๋ก ์๊ฐํ๊ณ ๋๋ก๋ ๋ ์ด์ ํ์ ์๋ ์์ ๋ณ์๋ฅผ ์ฆ์ ๋ช
์์ ์ผ๋ก ๋ฉ๋ชจ๋ฆฌ์์ ์ ๊ฑฐํ๋ ๊ฒ์ด ์ค์ํฉ๋๋ค.
**๊ธฐ๋ฅ๋ณ ๋ฉ๋ชจ๋ฆฌ**
๊ทธ๋ฐ ๋ค์, ์ํํธ์จ์ด์๋ ํน๋ณํ ๋ฉ๋ชจ๋ฆฌ ์๊ตฌ ์ฌํญ์ด ์์ ์ ์์ต๋๋ค. ์๋ฅผ ๋ค์ด, ๋น ๊ฒ์์ ์ฌ์ฉํ์ฌ ํ
์คํธ๋ฅผ ์์ฑํ ๋ ์ํํธ์จ์ด๋ ์
๋ ฅ๊ณผ ์ถ๋ ฅ ์ฌ๋ณธ์ ์ฌ๋ฌ ๊ฐ ์ ์งํด์ผ ํฉ๋๋ค.
**`forward` vs `backward` ์คํ ์๋**
ํฉ์ฑ๊ณฑ๊ณผ ์ ํ ๋ ์ด์ด์ ๊ฒฝ์ฐ ์๋ฐฉํฅ์ ๋นํด ์ญ๋ฐฉํฅ์์๋ 2๋ฐฐ์ ํ๋กญ์ค๊ฐ ํ์ํ๋ฏ๋ก ์ผ๋ฐ์ ์ผ๋ก 2๋ฐฐ ์ ๋ ๋๋ฆฌ๊ฒ ๋ณํ๋ฉ๋๋ค(์ญ๋ฐฉํฅ์ ๊ฒฝ์ฐ ์ฌ์ด์ฆ๊ฐ ๋ถ์์ฐ์ค๋ฝ๊ธฐ ๋๋ฌธ์, ๋๋ก๋ ๋์ฑ ๋๋ฆด ์๋ ์์ต๋๋ค). ํ์ฑํ๋ ์ผ๋ฐ์ ์ผ๋ก ๋์ญํญ์ด ์ ํ๋์ด ์์ผ๋ฉฐ, ์ผ๋ฐ์ ์ผ๋ก ์๋ฐฉํฅ๋ณด๋ค ์ญ๋ฐฉํฅ์์ ๋ ๋ง์ ๋ฐ์ดํฐ๋ฅผ ์ฝ์ด์ผ ํฉ๋๋ค. (์๋ฅผ ๋ค์ด, ์๋ฐฉํฅ ํ์ฑํ ์ ํ ๋ฒ ์ฉ ์ฝ๊ณ ์ฐ์ง๋ง, ์ญ๋ฐฉํฅ ํ์ฑํ์์๋ ์๋ฐฉํฅ gradOutput๊ณผ ์ถ๋ ฅ์ ๋ํด ์ด ๋ ๋ฒ ์ฝ๊ณ gradInput์ ๋ํด ํ ๋ฒ ์๋๋ค.)
๋ณด๋ค์ํผ, GPU ๋ฉ๋ชจ๋ฆฌ๋ฅผ ์ ์ฝํ๊ฑฐ๋ ์์
์๋๋ฅผ ๋์ผ ์ ์๋ ๋ช ๊ฐ์ง ๋ฐฉ๋ฒ์ด ์์ต๋๋ค.
์ด์ GPU ํ์ฉ๊ณผ ๊ณ์ฐ ์๋์ ์ํฅ์ ์ฃผ๋ ๊ฒ์ด ๋ฌด์์ธ์ง๋ฅผ ์ดํดํ์ผ๋ฏ๋ก, [Methods and tools for efficient training on a single GPU](perf_train_gpu_one) ๋ฌธ์ ํ์ด์ง๋ฅผ ์ฐธ์กฐํ์ฌ ์ฑ๋ฅ ์ต์ ํ ๊ธฐ๋ฒ์ ๋ํด ์์๋ณด์ธ์.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/model_sharing.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ๋ชจ๋ธ ๊ณต์ ํ๊ธฐ[[share-a-model]]
์ง๋ ๋ ํํ ๋ฆฌ์ผ์์ ๋ถ์ฐ ์ค์ ์ ์ํด PyTorch, Keras ๋ฐ ๐ค Accelerate๋ฅผ ์ฌ์ฉํ์ฌ ๋ชจ๋ธ์ ๋ฏธ์ธ ์กฐ์ ํ๋ ๋ฐฉ๋ฒ์ ๋ณด์์ต๋๋ค. ๋ค์ ๋จ๊ณ๋ ๋ชจ๋ธ์ ์ปค๋ฎค๋ํฐ์ ๊ณต์ ํ๋ ๊ฒ์
๋๋ค! Hugging Face๋ ์ธ๊ณต์ง๋ฅ์ ๋ฏผ์ฃผํ๋ฅผ ์ํด ๋ชจ๋์๊ฒ ์ง์๊ณผ ์์์ ๊ณต๊ฐ์ ์ผ๋ก ๊ณต์ ํด์ผ ํ๋ค๊ณ ๋ฏฟ์ต๋๋ค. ๋ค๋ฅธ ์ฌ๋๋ค์ด ์๊ฐ๊ณผ ์์์ ์ ์ฝํ ์ ์๋๋ก ์ปค๋ฎค๋ํฐ์ ๋ชจ๋ธ์ ๊ณต์ ํ๋ ๊ฒ์ ๊ณ ๋ คํด ๋ณด์ธ์.
์ด ํํ ๋ฆฌ์ผ์์ [Model Hub](https://huggingface.co/models)์์ ํ๋ จ๋๊ฑฐ๋ ๋ฏธ์ธ ์กฐ์ ๋ชจ๋ธ์ ๊ณต์ ํ๋ ๋ ๊ฐ์ง ๋ฐฉ๋ฒ์ ๋ํด ์์๋ด
์๋ค:
- API๋ฅผ ํตํด ํ์ผ์ Hub์ ํธ์ํฉ๋๋ค.
- ์น์ฌ์ดํธ๋ฅผ ํตํด ํ์ผ์ Hub๋ก ๋์ด๋ค ๋์ต๋๋ค.
<iframe width="560" height="315" src="https://www.youtube.com/embed/XvSGPZFEjDY" title="YouTube video player"
frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope;
picture-in-picture" allowfullscreen></iframe>
<Tip>
์ปค๋ฎค๋ํฐ์ ๋ชจ๋ธ์ ๊ณต์ ํ๋ ค๋ฉด, [huggingface.co](https://huggingface.co/join)์ ๊ณ์ ์ด ํ์ํฉ๋๋ค. ๊ธฐ์กด ์กฐ์ง์ ๊ฐ์
ํ๊ฑฐ๋ ์๋ก ๋ง๋ค ์๋ ์์ต๋๋ค.
</Tip>
## ์ ์ฅ์ ํน์ง[[repository-features]]
๋ชจ๋ธ ํ๋ธ์ ๊ฐ ์ ์ฅ์๋ ์ผ๋ฐ์ ์ธ GitHub ์ ์ฅ์์ฒ๋ผ ์๋ํฉ๋๋ค. ์ ์ฅ์๋ ๋ฒ์ ๊ด๋ฆฌ, ์ปค๋ฐ ๊ธฐ๋ก, ์ฐจ์ด์ ์๊ฐํ ๊ธฐ๋ฅ์ ์ ๊ณตํฉ๋๋ค.
๋ชจ๋ธ ํ๋ธ์ ๋ด์ฅ๋ ๋ฒ์ ๊ด๋ฆฌ๋ git ๋ฐ [git-lfs](https://git-lfs.github.com/)๋ฅผ ๊ธฐ๋ฐ์ผ๋ก ํฉ๋๋ค. ์ฆ, ํ๋์ ๋ชจ๋ธ์ ํ๋์ ์ ์ฅ์๋ก ์ทจ๊ธํ์ฌ ์ ๊ทผ ์ ์ด ๋ฐ ํ์ฅ์ฑ์ด ํฅ์๋ฉ๋๋ค. ๋ฒ์ ์ ์ด๋ ์ปค๋ฐ ํด์, ํ๊ทธ ๋๋ ๋ธ๋์น๋ก ๋ชจ๋ธ์ ํน์ ๋ฒ์ ์ ๊ณ ์ ํ๋ ๋ฐฉ๋ฒ์ธ *revision*์ ํ์ฉํฉ๋๋ค.
๋ฐ๋ผ์ `revision` ๋งค๊ฐ๋ณ์๋ฅผ ์ฌ์ฉํ์ฌ ํน์ ๋ชจ๋ธ ๋ฒ์ ์ ๊ฐ์ ธ์ฌ ์ ์์ต๋๋ค:
```py
>>> model = AutoModel.from_pretrained(
... "julien-c/EsperBERTo-small", revision="v2.0.1" # tag name, or branch name, or commit hash
... )
```
๋ํ ์ ์ฅ์์์ ํ์ผ์ ์ฝ๊ฒ ํธ์งํ ์ ์์ผ๋ฉฐ, ์ปค๋ฐ ๊ธฐ๋ก๊ณผ ์ฐจ์ด๋ฅผ ๋ณผ ์ ์์ต๋๋ค:
## ์ค์ [[setup]]
๋ชจ๋ธ์ ํ๋ธ์ ๊ณต์ ํ๊ธฐ ์ ์ Hugging Face ์๊ฒฉ ์ฆ๋ช
์ด ํ์ํฉ๋๋ค. ํฐ๋ฏธ๋์ ์ก์ธ์คํ ์ ์๋ ๊ฒฝ์ฐ, ๐ค Transformers๊ฐ ์ค์น๋ ๊ฐ์ ํ๊ฒฝ์์ ๋ค์ ๋ช
๋ น์ ์คํํฉ๋๋ค. ๊ทธ๋ฌ๋ฉด Hugging Face ์บ์ ํด๋(๊ธฐ๋ณธ์ ์ผ๋ก `~/.cache/`)์ ์ก์ธ์ค ํ ํฐ์ ์ ์ฅํฉ๋๋ค:
```bash
huggingface-cli login
```
Jupyter ๋๋ Colaboratory์ ๊ฐ์ ๋
ธํธ๋ถ์ ์ฌ์ฉ ์ค์ธ ๊ฒฝ์ฐ, [`huggingface_hub`](https://huggingface.co/docs/hub/adding-a-library) ๋ผ์ด๋ธ๋ฌ๋ฆฌ๊ฐ ์ค์น๋์๋์ง ํ์ธํ์ธ์. ์ด ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ์ฌ์ฉํ๋ฉด API๋ก ํ๋ธ์ ์ํธ ์์ฉํ ์ ์์ต๋๋ค.
```bash
pip install huggingface_hub
```
๊ทธ๋ฐ ๋ค์ `notebook_login`๋ก ํ๋ธ์ ๋ก๊ทธ์ธํ๊ณ , [์ฌ๊ธฐ](https://huggingface.co/settings/token) ๋งํฌ์์ ๋ก๊ทธ์ธํ ํ ํฐ์ ์์ฑํฉ๋๋ค:
```py
>>> from huggingface_hub import notebook_login
>>> notebook_login()
```
## ํ๋ ์์ํฌ ๊ฐ ๋ชจ๋ธ ๋ณํํ๊ธฐ[[convert-a-model-for-all-frameworks]]
๋ค๋ฅธ ํ๋ ์์ํฌ๋ก ์์
ํ๋ ์ฌ์ฉ์๊ฐ ๋ชจ๋ธ์ ์ฌ์ฉํ ์ ์๋๋ก ํ๋ ค๋ฉด, PyTorch ๋ฐ TensorFlow ์ฒดํฌํฌ์ธํธ๋ฅผ ๋ชจ๋ ์ฌ์ฉํ์ฌ ๋ชจ๋ธ์ ๋ณํํ๊ณ ์
๋ก๋ํ๋ ๊ฒ์ด ์ข์ต๋๋ค. ์ด ๋จ๊ณ๋ฅผ ๊ฑด๋๋ฐ์ด๋ ์ฌ์ฉ์๋ ๋ค๋ฅธ ํ๋ ์์ํฌ์์ ๋ชจ๋ธ์ ๊ฐ์ ธ์ฌ ์ ์์ง๋ง, ๐ค Transformers๊ฐ ์ฒดํฌํฌ์ธํธ๋ฅผ ์ฆ์์์ ๋ณํํด์ผ ํ๋ฏ๋ก ์๋๊ฐ ๋๋ ค์ง ์ ์์ต๋๋ค.
์ฒดํฌํฌ์ธํธ๋ฅผ ๋ค๋ฅธ ํ๋ ์์ํฌ๋ก ๋ณํํ๋ ๊ฒ์ ์ฝ์ต๋๋ค. PyTorch ๋ฐ TensorFlow๊ฐ ์ค์น๋์ด ์๋์ง ํ์ธํ ๋ค์(์ค์น ์ง์นจ์ [์ฌ๊ธฐ](installation) ์ฐธ์กฐ) ๋ค๋ฅธ ํ๋ ์์ํฌ์์ ์์
์ ๋ํ ํน์ ๋ชจ๋ธ์ ์ฐพ์ต๋๋ค.
<frameworkcontent>
<pt>
์ฒดํฌํฌ์ธํธ๋ฅผ TensorFlow์์ PyTorch๋ก ๋ณํํ๋ ค๋ฉด `from_tf=True`๋ฅผ ์ง์ ํ์ธ์:
```py
>>> pt_model = DistilBertForSequenceClassification.from_pretrained("path/to/awesome-name-you-picked", from_tf=True)
>>> pt_model.save_pretrained("path/to/awesome-name-you-picked")
```
</pt>
<tf>
์ฒดํฌํฌ์ธํธ๋ฅผ PyTorch์์ TensorFlow๋ก ๋ณํํ๋ ค๋ฉด `from_pt=True`๋ฅผ ์ง์ ํ์ธ์:
```py
>>> tf_model = TFDistilBertForSequenceClassification.from_pretrained("path/to/awesome-name-you-picked", from_pt=True)
```
๊ทธ๋ฐ ๋ค์ ์๋ก์ด ์ฒดํฌํฌ์ธํธ์ ํจ๊ป ์๋ก์ด TensorFlow ๋ชจ๋ธ์ ์ ์ฅํ ์ ์์ต๋๋ค:
```py
>>> tf_model.save_pretrained("path/to/awesome-name-you-picked")
```
</tf>
<jax>
Flax์์ ๋ชจ๋ธ์ ์ฌ์ฉํ๋ ๊ฒฝ์ฐ, PyTorch์์ Flax๋ก ์ฒดํฌํฌ์ธํธ๋ฅผ ๋ณํํ ์๋ ์์ต๋๋ค:
```py
>>> flax_model = FlaxDistilBertForSequenceClassification.from_pretrained(
... "path/to/awesome-name-you-picked", from_pt=True
... )
```
</jax>
</frameworkcontent>
## ํ๋ จ ์ค ๋ชจ๋ธ ํธ์ํ๊ธฐ[[push-a-model-during-training]]
<frameworkcontent>
<pt>
<Youtube id="Z1-XMy-GNLQ"/>
๋ชจ๋ธ์ ํ๋ธ์ ๊ณต์ ํ๋ ๊ฒ์ ์ถ๊ฐ ๋งค๊ฐ๋ณ์๋ ์ฝ๋ฐฑ์ ์ถ๊ฐํ๋ ๊ฒ๋งํผ ๊ฐ๋จํฉ๋๋ค. [๋ฏธ์ธ ์กฐ์ ํํ ๋ฆฌ์ผ](training)์์ [`TrainingArguments`] ํด๋์ค๋ ํ์ดํผํ๋ผ๋ฏธํฐ์ ์ถ๊ฐ ํ๋ จ ์ต์
์ ์ง์ ํ๋ ๊ณณ์ด๋ผ๋ ๊ฒ์ ๊ธฐ์ตํ์ธ์. ์ด๋ฌํ ํ๋ จ ์ต์
์ค ํ๋๋ ๋ชจ๋ธ์ ํ๋ธ๋ก ์ง์ ํธ์ํ๋ ๊ธฐ๋ฅ์ ํฌํจํฉ๋๋ค. [`TrainingArguments`]์์ `push_to_hub=True`๋ฅผ ์ค์ ํ์ธ์:
```py
>>> training_args = TrainingArguments(output_dir="my-awesome-model", push_to_hub=True)
```
ํ์์ ๊ฐ์ด ํ๋ จ ์ธ์๋ฅผ [`Trainer`]์ ์ ๋ฌํฉ๋๋ค:
```py
>>> trainer = Trainer(
... model=model,
... args=training_args,
... train_dataset=small_train_dataset,
... eval_dataset=small_eval_dataset,
... compute_metrics=compute_metrics,
... )
```
๋ชจ๋ธ์ ๋ฏธ์ธ ์กฐ์ ํ ํ, [`Trainer`]์์ [`~transformers.Trainer.push_to_hub`]๋ฅผ ํธ์ถํ์ฌ ํ๋ จ๋ ๋ชจ๋ธ์ ํ๋ธ๋ก ํธ์ํ์ธ์. ๐ค Transformers๋ ํ๋ จ ํ์ดํผํ๋ผ๋ฏธํฐ, ํ๋ จ ๊ฒฐ๊ณผ ๋ฐ ํ๋ ์์ํฌ ๋ฒ์ ์ ๋ชจ๋ธ ์นด๋์ ์๋์ผ๋ก ์ถ๊ฐํฉ๋๋ค!
```py
>>> trainer.push_to_hub()
```
</pt>
<tf>
[`PushToHubCallback`]์ ์ฌ์ฉํ์ฌ ๋ชจ๋ธ์ ํ๋ธ์ ๊ณต์ ํ๋ ค๋ฉด, [`PushToHubCallback`]์ ๋ค์ ์ธ์๋ฅผ ์ ์ํ์ธ์:
- ์ถ๋ ฅ๋ ๋ชจ๋ธ์ ํ์ผ ๊ฒฝ๋ก
- ํ ํฌ๋์ด์
- `{Hub ์ฌ์ฉ์ ์ด๋ฆ}/{๋ชจ๋ธ ์ด๋ฆ}` ํ์์ `hub_model_id`
```py
>>> from transformers import PushToHubCallback
>>> push_to_hub_callback = PushToHubCallback(
... output_dir="./your_model_save_path", tokenizer=tokenizer, hub_model_id="your-username/my-awesome-model"
... )
```
[`fit`](https://keras.io/api/models/model_training_apis/)์ ์ฝ๋ฐฑ์ ์ถ๊ฐํ๋ฉด, ๐ค Transformers๊ฐ ํ๋ จ๋ ๋ชจ๋ธ์ ํ๋ธ๋ก ํธ์ํฉ๋๋ค:
```py
>>> model.fit(tf_train_dataset, validation_data=tf_validation_dataset, epochs=3, callbacks=push_to_hub_callback)
```
</tf>
</frameworkcontent>
## `push_to_hub` ํจ์ ์ฌ์ฉํ๊ธฐ[[use-the-pushtohub-function]]
๋ชจ๋ธ์์ ์ง์ `push_to_hub`๋ฅผ ํธ์ถํ์ฌ ํ๋ธ์ ์
๋ก๋ํ ์๋ ์์ต๋๋ค.
`push_to_hub`์ ๋ชจ๋ธ ์ด๋ฆ์ ์ง์ ํ์ธ์:
```py
>>> pt_model.push_to_hub("my-awesome-model")
```
์ด๋ ๊ฒ ํ๋ฉด ์ฌ์ฉ์ ์ด๋ฆ ์๋์ ๋ชจ๋ธ ์ด๋ฆ `my-awesome-model`๋ก ์ ์ฅ์๊ฐ ์์ฑ๋ฉ๋๋ค. ์ด์ ์ฌ์ฉ์๋ `from_pretrained` ํจ์๋ฅผ ์ฌ์ฉํ์ฌ ๋ชจ๋ธ์ ๊ฐ์ ธ์ฌ ์ ์์ต๋๋ค:
```py
>>> from transformers import AutoModel
>>> model = AutoModel.from_pretrained("your_username/my-awesome-model")
```
์กฐ์ง์ ์ํ๊ณ ๋ชจ๋ธ์ ์กฐ์ง ์ด๋ฆ์ผ๋ก ๋์ ํธ์ํ๋ ค๋ฉด `repo_id`์ ์ถ๊ฐํ์ธ์:
```py
>>> pt_model.push_to_hub("my-awesome-org/my-awesome-model")
```
`push_to_hub` ํจ์๋ ๋ชจ๋ธ ์ ์ฅ์์ ๋ค๋ฅธ ํ์ผ์ ์ถ๊ฐํ๋ ๋ฐ์๋ ์ฌ์ฉํ ์ ์์ต๋๋ค. ์๋ฅผ ๋ค์ด ๋ชจ๋ธ ์ ์ฅ์์ ํ ํฌ๋์ด์ ๋ฅผ ์ถ๊ฐํ ์ ์์ต๋๋ค:
```py
>>> tokenizer.push_to_hub("my-awesome-model")
```
๋๋ ๋ฏธ์ธ ์กฐ์ ๋ PyTorch ๋ชจ๋ธ์ TensorFlow ๋ฒ์ ์ ์ถ๊ฐํ ์๋ ์์ต๋๋ค:
```py
>>> tf_model.push_to_hub("my-awesome-model")
```
์ด์ Hugging Face ํ๋กํ๋ก ์ด๋ํ๋ฉด, ์๋ก ์์ฑํ ๋ชจ๋ธ ์ ์ฅ์๊ฐ ํ์๋ฉ๋๋ค. **Files** ํญ์ ํด๋ฆญํ๋ฉด ์ ์ฅ์์ ์
๋ก๋ํ ๋ชจ๋ ํ์ผ์ด ํ์๋ฉ๋๋ค.
์ ์ฅ์์ ํ์ผ์ ๋ง๋ค๊ณ ์
๋ก๋ํ๋ ๋ฐฉ๋ฒ์ ๋ํ ์์ธํ ๋ด์ฉ์ ํ๋ธ ์ค๋ช
์ [์ฌ๊ธฐ](https://huggingface.co/docs/hub/how-to-upstream)๋ฅผ ์ฐธ์กฐํ์ธ์.
## ์น ์ธํฐํ์ด์ค๋ก ์
๋ก๋ํ๊ธฐ[[upload-with-the-web-interface]]
์ฝ๋ ์๋ ์ ๊ทผ ๋ฐฉ์์ ์ ํธํ๋ ์ฌ์ฉ์๋ ํ๋ธ์ ์น ์ธํฐํ์ด์ค๋ฅผ ํตํด ๋ชจ๋ธ์ ์
๋ก๋ํ ์ ์์ต๋๋ค. [huggingface.co/new](https://huggingface.co/new)๋ฅผ ๋ฐฉ๋ฌธํ์ฌ ์๋ก์ด ์ ์ฅ์๋ฅผ ์์ฑํ์ธ์:
์ฌ๊ธฐ์ ๋ชจ๋ธ์ ๋ํ ๋ช ๊ฐ์ง ์ ๋ณด๋ฅผ ์ถ๊ฐํ์ธ์:
- ์ ์ฅ์์ **์์ ์**๋ฅผ ์ ํํฉ๋๋ค. ์ด๋ ์ฌ์ฉ์ ๋๋ ์ฌ์ฉ์๊ฐ ์ํ ์กฐ์ง์ผ ์ ์์ต๋๋ค.
- ์ ์ฅ์ ์ด๋ฆ์ด ๋ ๋ชจ๋ธ์ ์ด๋ฆ์ ์ ํํฉ๋๋ค.
- ๋ชจ๋ธ์ด ๊ณต๊ฐ์ธ์ง ๋น๊ณต๊ฐ์ธ์ง ์ ํํฉ๋๋ค.
- ๋ชจ๋ธ์ ๋ผ์ด์ผ์ค ์ฌ์ฉ์ ์ง์ ํฉ๋๋ค.
์ด์ **Files** ํญ์ ํด๋ฆญํ๊ณ **Add file** ๋ฒํผ์ ํด๋ฆญํ์ฌ ์๋ก์ด ํ์ผ์ ์ ์ฅ์์ ์
๋ก๋ํฉ๋๋ค. ๊ทธ๋ฐ ๋ค์ ์
๋ก๋ํ ํ์ผ์ ๋์ด๋ค ๋๊ณ ์ปค๋ฐ ๋ฉ์์ง๋ฅผ ์ถ๊ฐํ์ธ์.
## ๋ชจ๋ธ ์นด๋ ์ถ๊ฐํ๊ธฐ[[add-a-model-card]]
์ฌ์ฉ์๊ฐ ๋ชจ๋ธ์ ๊ธฐ๋ฅ, ์ ํ, ์ ์ฌ์ ํธํฅ ๋ฐ ์ค๋ฆฌ์ ๊ณ ๋ ค ์ฌํญ์ ์ดํดํ ์ ์๋๋ก ์ ์ฅ์์ ๋ชจ๋ธ ์นด๋๋ฅผ ์ถ๊ฐํ์ธ์. ๋ชจ๋ธ ์นด๋๋ `README.md` ํ์ผ์ ์ ์๋์ด ์์ต๋๋ค. ๋ค์ ๋ฐฉ๋ฒ์ผ๋ก ๋ชจ๋ธ ์นด๋๋ฅผ ์ถ๊ฐํ ์ ์์ต๋๋ค:
* `README.md` ํ์ผ์ ์๋์ผ๋ก ์์ฑํ์ฌ ์
๋ก๋ํฉ๋๋ค.
* ๋ชจ๋ธ ์ ์ฅ์์์ **Edit model card** ๋ฒํผ์ ํด๋ฆญํฉ๋๋ค.
๋ชจ๋ธ ์นด๋์ ํฌํจํ ์ ๋ณด ์ ํ์ ๋ํ ์ข์ ์๋ DistilBert [๋ชจ๋ธ ์นด๋](https://huggingface.co/distilbert-base-uncased)๋ฅผ ์ฐธ์กฐํ์ธ์. ๋ชจ๋ธ์ ํ์ ๋ฐ์๊ตญ์ด๋ ์์ ฏ ์์ ๋ฑ `README.md` ํ์ผ์์ ์ ์ดํ ์ ์๋ ๋ค๋ฅธ ์ต์
์ ๋ํ ์์ธํ ๋ด์ฉ์ [์ฌ๊ธฐ](https://huggingface.co/docs/hub/models-cards) ๋ฌธ์๋ฅผ ์ฐธ์กฐํ์ธ์.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/create_a_model.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ๋ง์ถคํ ์ํคํ
์ฒ ๋ง๋ค๊ธฐ[[create-a-custom-architecture]]
[`AutoClass`](model_doc/auto)๋ ๋ชจ๋ธ ์ํคํ
์ฒ๋ฅผ ์๋์ผ๋ก ์ถ๋ก ํ๊ณ ๋ฏธ๋ฆฌ ํ์ต๋ configuration๊ณผ ๊ฐ์ค์น๋ฅผ ๋ค์ด๋ก๋ํฉ๋๋ค. ์ผ๋ฐ์ ์ผ๋ก ์ฒดํฌํฌ์ธํธ์ ๊ตฌ์ ๋ฐ์ง ์๋ ์ฝ๋๋ฅผ ์์ฑํ๋ ค๋ฉด `AutoClass`๋ฅผ ์ฌ์ฉํ๋ ๊ฒ์ด ์ข์ต๋๋ค. ํ์ง๋ง ํน์ ๋ชจ๋ธ ํ๋ผ๋ฏธํฐ๋ฅผ ๋ณด๋ค ์ธ๋ฐํ๊ฒ ์ ์ดํ๊ณ ์ ํ๋ ์ฌ์ฉ์๋ ๋ช ๊ฐ์ง ๊ธฐ๋ณธ ํด๋์ค๋ง์ผ๋ก ์ปค์คํ
๐ค Transformers ๋ชจ๋ธ์ ์์ฑํ ์ ์์ต๋๋ค. ์ด๋ ๐ค Transformers ๋ชจ๋ธ์ ์ฐ๊ตฌ, ๊ต์ก ๋๋ ์คํํ๋ ๋ฐ ๊ด์ฌ์ด ์๋ ๋ชจ๋ ์ฌ์ฉ์์๊ฒ ํนํ ์ ์ฉํ ์ ์์ต๋๋ค. ์ด ๊ฐ์ด๋์์๋ 'AutoClass'๋ฅผ ์ฌ์ฉํ์ง ์๊ณ ์ปค์คํ
๋ชจ๋ธ์ ๋ง๋๋ ๋ฐฉ๋ฒ์ ๋ํด ์์๋ณด๊ฒ ์ต๋๋ค:
- ๋ชจ๋ธ configuration์ ๊ฐ์ ธ์ค๊ณ ์ฌ์ฉ์ ์ง์ ํฉ๋๋ค.
- ๋ชจ๋ธ ์ํคํ
์ฒ๋ฅผ ์์ฑํฉ๋๋ค.
- ํ
์คํธ์ ์ฌ์ฉํ ๋๋ฆฌ๊ฑฐ๋ ๋น ๋ฅธ ํ ํฐํ๊ธฐ๋ฅผ ๋ง๋ญ๋๋ค.
- ๋น์ ์์
์ ์ํ ์ด๋ฏธ์ง ํ๋ก์ธ์๋ฅผ ์์ฑํฉ๋๋ค.
- ์ค๋์ค ์์
์ ์ํ ํน์ฑ ์ถ์ถ๊ธฐ๋ฅผ ์์ฑํฉ๋๋ค.
- ๋ฉํฐ๋ชจ๋ฌ ์์
์ฉ ํ๋ก์ธ์๋ฅผ ์์ฑํฉ๋๋ค.
## Configuration[[configuration]]
[configuration](main_classes/configuration)์ ๋ชจ๋ธ์ ํน์ ์์ฑ์ ๋ํ๋
๋๋ค. ๊ฐ ๋ชจ๋ธ ๊ตฌ์ฑ์๋ ์๋ก ๋ค๋ฅธ ์์ฑ์ด ์์ต๋๋ค. ์๋ฅผ ๋ค์ด, ๋ชจ๋ NLP ๋ชจ๋ธ์๋ `hidden_size`, `num_attention_heads`, `num_hidden_layers` ๋ฐ `vocab_size` ์์ฑ์ด ๊ณตํต์ผ๋ก ์์ต๋๋ค. ์ด๋ฌํ ์์ฑ์ ๋ชจ๋ธ์ ๊ตฌ์ฑํ attention heads ๋๋ hidden layers์ ์๋ฅผ ์ง์ ํฉ๋๋ค.
[DistilBERT](model_doc/distilbert) ์์ฑ์ ๊ฒ์ฌํ๊ธฐ ์ํด [`DistilBertConfig`]์ ์ ๊ทผํ์ฌ ์์ธํ ์ดํด๋ด
๋๋ค:
```py
>>> from transformers import DistilBertConfig
>>> config = DistilBertConfig()
>>> print(config)
DistilBertConfig {
"activation": "gelu",
"attention_dropout": 0.1,
"dim": 768,
"dropout": 0.1,
"hidden_dim": 3072,
"initializer_range": 0.02,
"max_position_embeddings": 512,
"model_type": "distilbert",
"n_heads": 12,
"n_layers": 6,
"pad_token_id": 0,
"qa_dropout": 0.1,
"seq_classif_dropout": 0.2,
"sinusoidal_pos_embds": false,
"transformers_version": "4.16.2",
"vocab_size": 30522
}
```
[`DistilBertConfig`]๋ ๊ธฐ๋ณธ [`DistilBertModel`]์ ๋น๋ํ๋ ๋ฐ ์ฌ์ฉ๋๋ ๋ชจ๋ ๊ธฐ๋ณธ ์์ฑ์ ํ์ํฉ๋๋ค. ๋ชจ๋ ์์ฑ์ ์ปค์คํฐ๋ง์ด์ง์ด ๊ฐ๋ฅํ๋ฏ๋ก ์คํ์ ์ํ ๊ณต๊ฐ์ ๋ง๋ค ์ ์์ต๋๋ค. ์๋ฅผ ๋ค์ด ๊ธฐ๋ณธ ๋ชจ๋ธ์ ๋ค์๊ณผ ๊ฐ์ด ์ปค์คํฐ๋ง์ด์ฆํ ์ ์์ต๋๋ค:
- `activation` ํ๋ผ๋ฏธํฐ๋ก ๋ค๋ฅธ ํ์ฑํ ํจ์๋ฅผ ์ฌ์ฉํด ๋ณด์ธ์.
- `attention_dropout` ํ๋ผ๋ฏธํฐ๋ฅผ ์ฌ์ฉํ์ฌ ์ดํ
์
ํ๋ฅ ์ ๋ ๋์ ๋๋กญ์์ ๋น์จ์ ์ฌ์ฉํ์ธ์.
```py
>>> my_config = DistilBertConfig(activation="relu", attention_dropout=0.4)
>>> print(my_config)
DistilBertConfig {
"activation": "relu",
"attention_dropout": 0.4,
"dim": 768,
"dropout": 0.1,
"hidden_dim": 3072,
"initializer_range": 0.02,
"max_position_embeddings": 512,
"model_type": "distilbert",
"n_heads": 12,
"n_layers": 6,
"pad_token_id": 0,
"qa_dropout": 0.1,
"seq_classif_dropout": 0.2,
"sinusoidal_pos_embds": false,
"transformers_version": "4.16.2",
"vocab_size": 30522
}
```
์ฌ์ ํ์ต๋ ๋ชจ๋ธ ์์ฑ์ [`~PretrainedConfig.from_pretrained`] ํจ์์์ ์์ ํ ์ ์์ต๋๋ค:
```py
>>> my_config = DistilBertConfig.from_pretrained("distilbert-base-uncased", activation="relu", attention_dropout=0.4)
```
๋ชจ๋ธ ๊ตฌ์ฑ์ด ๋ง์กฑ์ค๋ฌ์ฐ๋ฉด [`~PretrainedConfig.save_pretrained`]๋ก ์ ์ฅํ ์ ์์ต๋๋ค. ์ค์ ํ์ผ์ ์ง์ ๋ ์์
๊ฒฝ๋ก์ JSON ํ์ผ๋ก ์ ์ฅ๋ฉ๋๋ค:
```py
>>> my_config.save_pretrained(save_directory="./your_model_save_path")
```
configuration ํ์ผ์ ์ฌ์ฌ์ฉํ๋ ค๋ฉด [`~PretrainedConfig.from_pretrained`]๋ฅผ ์ฌ์ฉํ์ฌ ๊ฐ์ ธ์ค์ธ์:
```py
>>> my_config = DistilBertConfig.from_pretrained("./your_model_save_path/config.json")
```
<Tip>
configuration ํ์ผ์ ๋์
๋๋ฆฌ๋ก ์ ์ฅํ๊ฑฐ๋ ์ฌ์ฉ์ ์ ์ configuration ์์ฑ๊ณผ ๊ธฐ๋ณธ configuration ์์ฑ์ ์ฐจ์ด์ ๋ง ์ ์ฅํ ์๋ ์์ต๋๋ค! ์์ธํ ๋ด์ฉ์ [configuration](main_classes/configuration) ๋ฌธ์๋ฅผ ์ฐธ์กฐํ์ธ์.
</Tip>
## ๋ชจ๋ธ[[model]]
๋ค์ ๋จ๊ณ๋ [๋ชจ๋ธ(model)](main_classes/models)์ ๋ง๋๋ ๊ฒ์
๋๋ค. ๋์จํ๊ฒ ์ํคํ
์ฒ๋ผ๊ณ ๋ ๋ถ๋ฆฌ๋ ๋ชจ๋ธ์ ๊ฐ ๊ณ์ธต์ด ์ํํ๋ ๋์๊ณผ ๋ฐ์ํ๋ ์์
์ ์ ์ํฉ๋๋ค. configuration์ `num_hidden_layers`์ ๊ฐ์ ์์ฑ์ ์ํคํ
์ฒ๋ฅผ ์ ์ํ๋ ๋ฐ ์ฌ์ฉ๋ฉ๋๋ค. ๋ชจ๋ ๋ชจ๋ธ์ ๊ธฐ๋ณธ ํด๋์ค [`PreTrainedModel`]๊ณผ ์
๋ ฅ ์๋ฒ ๋ฉ ํฌ๊ธฐ ์กฐ์ ๋ฐ ์
ํ ์ดํ
์
ํค๋ ๊ฐ์ง ์น๊ธฐ์ ๊ฐ์ ๋ช ๊ฐ์ง ์ผ๋ฐ์ ์ธ ๋ฉ์๋๋ฅผ ๊ณต์ ํฉ๋๋ค. ๋ํ ๋ชจ๋ ๋ชจ๋ธ์ [`torch.nn.Module`](https://pytorch.org/docs/stable/generated/torch.nn.Module.html), [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) ๋๋ [`flax.linen.Module`](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html)์ ์๋ธํด๋์ค์ด๊ธฐ๋ ํฉ๋๋ค. ์ฆ, ๋ชจ๋ธ์ ๊ฐ ํ๋ ์์ํฌ์ ์ฌ์ฉ๋ฒ๊ณผ ํธํ๋ฉ๋๋ค.
<frameworkcontent>
<pt>
์ฌ์ฉ์ ์ง์ configuration ์์ฑ์ ๋ชจ๋ธ์ ๊ฐ์ ธ์ต๋๋ค:
```py
>>> from transformers import DistilBertModel
>>> my_config = DistilBertConfig.from_pretrained("./your_model_save_path/config.json")
>>> model = DistilBertModel(my_config)
```
์ด์ ์ฌ์ ํ์ต๋ ๊ฐ์ค์น ๋์ ์์์ ๊ฐ์ ๊ฐ์ง ๋ชจ๋ธ์ด ์์ฑ๋ฉ๋๋ค. ์ด ๋ชจ๋ธ์ ํ๋ จํ๊ธฐ ์ ๊น์ง๋ ์ ์ฉํ๊ฒ ์ฌ์ฉํ ์ ์์ต๋๋ค. ํ๋ จ์ ๋น์ฉ๊ณผ ์๊ฐ์ด ๋ง์ด ์์๋๋ ํ๋ก์ธ์ค์
๋๋ค. ์ผ๋ฐ์ ์ผ๋ก ํ๋ จ์ ํ์ํ ๋ฆฌ์์ค์ ์ผ๋ถ๋ง ์ฌ์ฉํ๋ฉด์ ๋ ๋์ ๊ฒฐ๊ณผ๋ฅผ ๋ ๋นจ๋ฆฌ ์ป์ผ๋ ค๋ฉด ์ฌ์ ํ๋ จ๋ ๋ชจ๋ธ์ ์ฌ์ฉํ๋ ๊ฒ์ด ์ข์ต๋๋ค.
์ฌ์ ํ์ต๋ ๋ชจ๋ธ์ [`~PreTrainedModel.from_pretrained`]๋ก ์์ฑํฉ๋๋ค:
```py
>>> model = DistilBertModel.from_pretrained("distilbert-base-uncased")
```
๐ค Transformers์์ ์ ๊ณตํ ๋ชจ๋ธ์ ์ฌ์ ํ์ต๋ ๊ฐ์ค์น๋ฅผ ์ฌ์ฉํ๋ ๊ฒฝ์ฐ ๊ธฐ๋ณธ ๋ชจ๋ธ configuration์ ์๋์ผ๋ก ๋ถ๋ฌ์ต๋๋ค. ๊ทธ๋ฌ๋ ์ํ๋ ๊ฒฝ์ฐ ๊ธฐ๋ณธ ๋ชจ๋ธ configuration ์์ฑ์ ์ผ๋ถ ๋๋ ์ ๋ถ๋ฅผ ์ฌ์ฉ์ ์ง์ ์ผ๋ก ๋ฐ๊ฟ ์ ์์ต๋๋ค:
```py
>>> model = DistilBertModel.from_pretrained("distilbert-base-uncased", config=my_config)
```
</pt>
<tf>
์ฌ์ฉ์ ์ง์ configuration ์์ฑ์ ๋ชจ๋ธ์ ๋ถ๋ฌ์ต๋๋ค:
```py
>>> from transformers import TFDistilBertModel
>>> my_config = DistilBertConfig.from_pretrained("./your_model_save_path/my_config.json")
>>> tf_model = TFDistilBertModel(my_config)
```
์ด์ ์ฌ์ ํ์ต๋ ๊ฐ์ค์น ๋์ ์์์ ๊ฐ์ ๊ฐ์ง ๋ชจ๋ธ์ด ์์ฑ๋ฉ๋๋ค. ์ด ๋ชจ๋ธ์ ํ๋ จํ๊ธฐ ์ ๊น์ง๋ ์ ์ฉํ๊ฒ ์ฌ์ฉํ ์ ์์ต๋๋ค. ํ๋ จ์ ๋น์ฉ๊ณผ ์๊ฐ์ด ๋ง์ด ์์๋๋ ํ๋ก์ธ์ค์
๋๋ค. ์ผ๋ฐ์ ์ผ๋ก ํ๋ จ์ ํ์ํ ๋ฆฌ์์ค์ ์ผ๋ถ๋ง ์ฌ์ฉํ๋ฉด์ ๋ ๋์ ๊ฒฐ๊ณผ๋ฅผ ๋ ๋นจ๋ฆฌ ์ป์ผ๋ ค๋ฉด ์ฌ์ ํ๋ จ๋ ๋ชจ๋ธ์ ์ฌ์ฉํ๋ ๊ฒ์ด ์ข์ต๋๋ค.
์ฌ์ ํ์ต๋ ๋ชจ๋ธ์ [`~TFPreTrainedModel.from_pretrained`]๋ก ์์ฑํฉ๋๋ค:
```py
>>> tf_model = TFDistilBertModel.from_pretrained("distilbert-base-uncased")
```
๐ค Transformers์์ ์ ๊ณตํ ๋ชจ๋ธ์ ์ฌ์ ํ์ต๋ ๊ฐ์ค์น๋ฅผ ์ฌ์ฉํ๋ ๊ฒฝ์ฐ ๊ธฐ๋ณธ ๋ชจ๋ธ configuration์ ์๋์ผ๋ก ๋ถ๋ฌ์ต๋๋ค. ๊ทธ๋ฌ๋ ์ํ๋ ๊ฒฝ์ฐ ๊ธฐ๋ณธ ๋ชจ๋ธ configuration ์์ฑ์ ์ผ๋ถ ๋๋ ์ ๋ถ๋ฅผ ์ฌ์ฉ์ ์ง์ ์ผ๋ก ๋ฐ๊ฟ ์ ์์ต๋๋ค:
```py
>>> tf_model = TFDistilBertModel.from_pretrained("distilbert-base-uncased", config=my_config)
```
</tf>
</frameworkcontent>
### ๋ชจ๋ธ ํค๋[[model-heads]]
์ด ์์ ์์ *์๋ ์ํ(hidden state)*๋ฅผ ์ถ๋ ฅํ๋ ๊ธฐ๋ณธ DistilBERT ๋ชจ๋ธ์ ๊ฐ๊ฒ ๋ฉ๋๋ค. ์๋ ์ํ๋ ์ต์ข
์ถ๋ ฅ์ ์์ฑํ๊ธฐ ์ํด ๋ชจ๋ธ ํค๋์ ์
๋ ฅ์ผ๋ก ์ ๋ฌ๋ฉ๋๋ค. ๐ค Transformers๋ ๋ชจ๋ธ์ด ํด๋น ์์
์ ์ง์ํ๋ ํ ๊ฐ ์์
๋ง๋ค ๋ค๋ฅธ ๋ชจ๋ธ ํค๋๋ฅผ ์ ๊ณตํฉ๋๋ค(์ฆ, ๋ฒ์ญ๊ณผ ๊ฐ์ ์ํ์ค ๊ฐ ์์
์๋ DistilBERT๋ฅผ ์ฌ์ฉํ ์ ์์).
<frameworkcontent>
<pt>
์๋ฅผ ๋ค์ด, [`DistilBertForSequenceClassification`]์ ์ํ์ค ๋ถ๋ฅ ํค๋๊ฐ ์๋ ๊ธฐ๋ณธ DistilBERT ๋ชจ๋ธ์
๋๋ค. ์ํ์ค ๋ถ๋ฅ ํค๋๋ ํ๋ง๋ ์ถ๋ ฅ ์์ ์๋ ์ ํ ๋ ์ด์ด์
๋๋ค.
```py
>>> from transformers import DistilBertForSequenceClassification
>>> model = DistilBertForSequenceClassification.from_pretrained("distilbert-base-uncased")
```
๋ค๋ฅธ ๋ชจ๋ธ ํค๋๋ก ์ ํํ์ฌ ์ด ์ฒดํฌํฌ์ธํธ๋ฅผ ๋ค๋ฅธ ์์
์ ์ฝ๊ฒ ์ฌ์ฌ์ฉํ ์ ์์ต๋๋ค. ์ง์์๋ต ์์
์ ๊ฒฝ์ฐ, [`DistilBertForQuestionAnswering`] ๋ชจ๋ธ ํค๋๋ฅผ ์ฌ์ฉํ ์ ์์ต๋๋ค. ์ง์์๋ต ํค๋๋ ์จ๊ฒจ์ง ์ํ ์ถ๋ ฅ ์์ ์ ํ ๋ ์ด์ด๊ฐ ์๋ค๋ ์ ์ ์ ์ธํ๋ฉด ์ํ์ค ๋ถ๋ฅ ํค๋์ ์ ์ฌํฉ๋๋ค.
```py
>>> from transformers import DistilBertForQuestionAnswering
>>> model = DistilBertForQuestionAnswering.from_pretrained("distilbert-base-uncased")
```
</pt>
<tf>
์๋ฅผ ๋ค์ด, [`TFDistilBertForSequenceClassification`]์ ์ํ์ค ๋ถ๋ฅ ํค๋๊ฐ ์๋ ๊ธฐ๋ณธ DistilBERT ๋ชจ๋ธ์
๋๋ค. ์ํ์ค ๋ถ๋ฅ ํค๋๋ ํ๋ง๋ ์ถ๋ ฅ ์์ ์๋ ์ ํ ๋ ์ด์ด์
๋๋ค.
```py
>>> from transformers import TFDistilBertForSequenceClassification
>>> tf_model = TFDistilBertForSequenceClassification.from_pretrained("distilbert-base-uncased")
```
๋ค๋ฅธ ๋ชจ๋ธ ํค๋๋ก ์ ํํ์ฌ ์ด ์ฒดํฌํฌ์ธํธ๋ฅผ ๋ค๋ฅธ ์์
์ ์ฝ๊ฒ ์ฌ์ฌ์ฉํ ์ ์์ต๋๋ค. ์ง์์๋ต ์์
์ ๊ฒฝ์ฐ, [`TFDistilBertForQuestionAnswering`] ๋ชจ๋ธ ํค๋๋ฅผ ์ฌ์ฉํ ์ ์์ต๋๋ค. ์ง์์๋ต ํค๋๋ ์จ๊ฒจ์ง ์ํ ์ถ๋ ฅ ์์ ์ ํ ๋ ์ด์ด๊ฐ ์๋ค๋ ์ ์ ์ ์ธํ๋ฉด ์ํ์ค ๋ถ๋ฅ ํค๋์ ์ ์ฌํฉ๋๋ค.
```py
>>> from transformers import TFDistilBertForQuestionAnswering
>>> tf_model = TFDistilBertForQuestionAnswering.from_pretrained("distilbert-base-uncased")
```
</tf>
</frameworkcontent>
## ํ ํฌ๋์ด์ [[tokenizer]]
ํ
์คํธ ๋ฐ์ดํฐ์ ๋ชจ๋ธ์ ์ฌ์ฉํ๊ธฐ ์ ์ ๋ง์ง๋ง์ผ๋ก ํ์ํ ๊ธฐ๋ณธ ํด๋์ค๋ ์์ ํ
์คํธ๋ฅผ ํ
์๋ก ๋ณํํ๋ [ํ ํฌ๋์ด์ ](main_classes/tokenizer)์
๋๋ค. ๐ค Transformers์ ์ฌ์ฉํ ์ ์๋ ํ ํฌ๋์ด์ ๋ ๋ ๊ฐ์ง ์ ํ์ด ์์ต๋๋ค:
- [`PreTrainedTokenizer`]: ํ์ด์ฌ์ผ๋ก ๊ตฌํ๋ ํ ํฌ๋์ด์ ์
๋๋ค.
- [`PreTrainedTokenizerFast`]: Rust ๊ธฐ๋ฐ [๐ค Tokenizer](https://huggingface.co/docs/tokenizers/python/latest/) ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ก ๋ง๋ค์ด์ง ํ ํฌ๋์ด์ ์
๋๋ค. ์ด ํ ํฌ๋์ด์ ๋ Rust๋ก ๊ตฌํ๋์ด ๋ฐฐ์น ํ ํฐํ์์ ํนํ ๋น ๋ฆ
๋๋ค. ๋น ๋ฅธ ํ ํฌ๋์ด์ ๋ ํ ํฐ์ ์๋ ๋จ์ด๋ ๋ฌธ์์ ๋งคํํ๋ *์คํ์
๋งคํ*๊ณผ ๊ฐ์ ์ถ๊ฐ ๋ฉ์๋๋ ์ ๊ณตํฉ๋๋ค.
๋ ํ ํฌ๋์ด์ ๋ชจ๋ ์ธ์ฝ๋ฉ ๋ฐ ๋์ฝ๋ฉ, ์ ํ ํฐ ์ถ๊ฐ, ํน์ ํ ํฐ ๊ด๋ฆฌ์ ๊ฐ์ ์ผ๋ฐ์ ์ธ ๋ฐฉ๋ฒ์ ์ง์ํฉ๋๋ค.
<Tip warning={true}>
๋ชจ๋ ๋ชจ๋ธ์ด ๋น ๋ฅธ ํ ํฌ๋์ด์ ๋ฅผ ์ง์ํ๋ ๊ฒ์ ์๋๋๋ค. ์ด [ํ](index#supported-frameworks)์์ ๋ชจ๋ธ์ ๋น ๋ฅธ ํ ํฌ๋์ด์ ์ง์ ์ฌ๋ถ๋ฅผ ํ์ธํ์ธ์.
</Tip>
ํ ํฌ๋์ด์ ๋ฅผ ์ง์ ํ์ตํ ๊ฒฝ์ฐ, *์ดํ(vocabulary)* ํ์ผ์์ ํ ํฌ๋์ด์ ๋ฅผ ๋ง๋ค ์ ์์ต๋๋ค:
```py
>>> from transformers import DistilBertTokenizer
>>> my_tokenizer = DistilBertTokenizer(vocab_file="my_vocab_file.txt", do_lower_case=False, padding_side="left")
```
์ฌ์ฉ์ ์ง์ ํ ํฌ๋์ด์ ์ ์ดํ๋ ์ฌ์ ํ์ต๋ ๋ชจ๋ธ์ ํ ํฌ๋์ด์ ์์ ์์ฑ๋ ์ดํ์ ๋ค๋ฅผ ์ ์๋ค๋ ์ ์ ๊ธฐ์ตํ๋ ๊ฒ์ด ์ค์ํฉ๋๋ค. ์ฌ์ ํ์ต๋ ๋ชจ๋ธ์ ์ฌ์ฉํ๋ ๊ฒฝ์ฐ ์ฌ์ ํ์ต๋ ๋ชจ๋ธ์ ์ดํ๋ฅผ ์ฌ์ฉํด์ผ ํ๋ฉฐ, ๊ทธ๋ ์ง ์์ผ๋ฉด ์
๋ ฅ์ด ์๋ฏธ๋ฅผ ๊ฐ์ง ๋ชปํฉ๋๋ค. [`DistilBertTokenizer`] ํด๋์ค๋ฅผ ์ฌ์ฉํ์ฌ ์ฌ์ ํ์ต๋ ๋ชจ๋ธ์ ์ดํ๋ก ํ ํฌ๋์ด์ ๋ฅผ ์์ฑํฉ๋๋ค:
```py
>>> from transformers import DistilBertTokenizer
>>> slow_tokenizer = DistilBertTokenizer.from_pretrained("distilbert-base-uncased")
```
[`DistilBertTokenizerFast`] ํด๋์ค๋ก ๋น ๋ฅธ ํ ํฌ๋์ด์ ๋ฅผ ์์ฑํฉ๋๋ค:
```py
>>> from transformers import DistilBertTokenizerFast
>>> fast_tokenizer = DistilBertTokenizerFast.from_pretrained("distilbert-base-uncased")
```
<Tip>
[`AutoTokenizer`]๋ ๊ธฐ๋ณธ์ ์ผ๋ก ๋น ๋ฅธ ํ ํฌ๋์ด์ ๋ฅผ ๊ฐ์ ธ์ค๋ ค๊ณ ํฉ๋๋ค. ์ด ๋์์ ๋นํ์ฑํํ๋ ค๋ฉด `from_pretrained`์์ `use_fast=False`๋ฅผ ์ค์ ํ๋ฉด ๋ฉ๋๋ค.
</Tip>
## ์ด๋ฏธ์ง ํ๋ก์ธ์[[image-processor]]
์ด๋ฏธ์ง ํ๋ก์ธ์(image processor)๋ ๋น์ ์
๋ ฅ์ ์ฒ๋ฆฌํฉ๋๋ค. ๊ธฐ๋ณธ [`~image_processing_utils.ImageProcessingMixin`] ํด๋์ค์์ ์์ํฉ๋๋ค.
์ฌ์ฉํ๋ ค๋ฉด ์ฌ์ฉ ์ค์ธ ๋ชจ๋ธ๊ณผ ์ฐ๊ฒฐ๋ ์ด๋ฏธ์ง ํ๋ก์ธ์๋ฅผ ์์ฑํฉ๋๋ค. ์๋ฅผ ๋ค์ด, ์ด๋ฏธ์ง ๋ถ๋ฅ์ [ViT](model_doc/vit)๋ฅผ ์ฌ์ฉํ๋ ๊ฒฝ์ฐ ๊ธฐ๋ณธ [`ViTImageProcessor`]๋ฅผ ์์ฑํฉ๋๋ค:
```py
>>> from transformers import ViTImageProcessor
>>> vit_extractor = ViTImageProcessor()
>>> print(vit_extractor)
ViTImageProcessor {
"do_normalize": true,
"do_resize": true,
"feature_extractor_type": "ViTImageProcessor",
"image_mean": [
0.5,
0.5,
0.5
],
"image_std": [
0.5,
0.5,
0.5
],
"resample": 2,
"size": 224
}
```
<Tip>
์ฌ์ฉ์ ์ง์ ์ ์ํ์ง ์๋ ๊ฒฝ์ฐ `from_pretrained` ๋ฉ์๋๋ฅผ ์ฌ์ฉํ์ฌ ๋ชจ๋ธ์ ๊ธฐ๋ณธ ์ด๋ฏธ์ง ํ๋ก์ธ์ ๋งค๊ฐ๋ณ์๋ฅผ ๋ถ๋ฌ์ค๋ฉด ๋ฉ๋๋ค.
</Tip>
์ฌ์ฉ์ ์ง์ ์ด๋ฏธ์ง ํ๋ก์ธ์๋ฅผ ์์ฑํ๋ ค๋ฉด [`ViTImageProcessor`] ํ๋ผ๋ฏธํฐ๋ฅผ ์์ ํฉ๋๋ค:
```py
>>> from transformers import ViTImageProcessor
>>> my_vit_extractor = ViTImageProcessor(resample="PIL.Image.BOX", do_normalize=False, image_mean=[0.3, 0.3, 0.3])
>>> print(my_vit_extractor)
ViTImageProcessor {
"do_normalize": false,
"do_resize": true,
"feature_extractor_type": "ViTImageProcessor",
"image_mean": [
0.3,
0.3,
0.3
],
"image_std": [
0.5,
0.5,
0.5
],
"resample": "PIL.Image.BOX",
"size": 224
}
```
## ํน์ฑ ์ถ์ถ๊ธฐ[[feature-extractor]]
ํน์ฑ ์ถ์ถ๊ธฐ(feature extractor)๋ ์ค๋์ค ์
๋ ฅ์ ์ฒ๋ฆฌํฉ๋๋ค. ๊ธฐ๋ณธ [`~feature_extraction_utils.FeatureExtractionMixin`] ํด๋์ค์์ ์์๋๋ฉฐ, ์ค๋์ค ์
๋ ฅ์ ์ฒ๋ฆฌํ๊ธฐ ์ํด [`SequenceFeatureExtractor`] ํด๋์ค์์ ์์ํ ์๋ ์์ต๋๋ค.
์ฌ์ฉํ๋ ค๋ฉด ์ฌ์ฉ ์ค์ธ ๋ชจ๋ธ๊ณผ ์ฐ๊ฒฐ๋ ํน์ฑ ์ถ์ถ๊ธฐ๋ฅผ ์์ฑํฉ๋๋ค. ์๋ฅผ ๋ค์ด, ์ค๋์ค ๋ถ๋ฅ์ [Wav2Vec2](model_doc/wav2vec2)๋ฅผ ์ฌ์ฉํ๋ ๊ฒฝ์ฐ ๊ธฐ๋ณธ [`Wav2Vec2FeatureExtractor`]๋ฅผ ์์ฑํฉ๋๋ค:
```py
>>> from transformers import Wav2Vec2FeatureExtractor
>>> w2v2_extractor = Wav2Vec2FeatureExtractor()
>>> print(w2v2_extractor)
Wav2Vec2FeatureExtractor {
"do_normalize": true,
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
"feature_size": 1,
"padding_side": "right",
"padding_value": 0.0,
"return_attention_mask": false,
"sampling_rate": 16000
}
```
<Tip>
์ฌ์ฉ์ ์ง์ ์ด ํ์ํ์ง ์์ ๊ฒฝ์ฐ `from_pretrained` ๋ฉ์๋๋ฅผ ์ฌ์ฉํ์ฌ ๋ชจ๋ธ์ ๊ธฐ๋ณธ ํน์ฑ ์ถ์ถ๊ธฐ ใ
๊ฐ๋ณ์๋ฅผ ๋ถ๋ฌ ์ค๋ฉด ๋ฉ๋๋ค.
</Tip>
์ฌ์ฉ์ ์ง์ ํน์ฑ ์ถ์ถ๊ธฐ๋ฅผ ๋ง๋ค๋ ค๋ฉด [`Wav2Vec2FeatureExtractor`] ๋งค๊ฐ๋ณ์๋ฅผ ์์ ํฉ๋๋ค:
```py
>>> from transformers import Wav2Vec2FeatureExtractor
>>> w2v2_extractor = Wav2Vec2FeatureExtractor(sampling_rate=8000, do_normalize=False)
>>> print(w2v2_extractor)
Wav2Vec2FeatureExtractor {
"do_normalize": false,
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
"feature_size": 1,
"padding_side": "right",
"padding_value": 0.0,
"return_attention_mask": false,
"sampling_rate": 8000
}
```
## ํ๋ก์ธ์[[processor]]
๋ฉํฐ๋ชจ๋ฌ ์์
์ ์ง์ํ๋ ๋ชจ๋ธ์ ๊ฒฝ์ฐ, ๐ค Transformers๋ ํน์ฑ ์ถ์ถ๊ธฐ ๋ฐ ํ ํฌ๋์ด์ ์ ๊ฐ์ ์ฒ๋ฆฌ ํด๋์ค๋ฅผ ๋จ์ผ ๊ฐ์ฒด๋ก ํธ๋ฆฌํ๊ฒ ๋ํํ๋ ํ๋ก์ธ์ ํด๋์ค๋ฅผ ์ ๊ณตํฉ๋๋ค. ์๋ฅผ ๋ค์ด, ์๋ ์์ฑ ์ธ์ ์์
(Automatic Speech Recognition task (ASR))์ [`Wav2Vec2Processor`]๋ฅผ ์ฌ์ฉํ๋ค๊ณ ๊ฐ์ ํด ๋ณด๊ฒ ์ต๋๋ค. ์๋ ์์ฑ ์ธ์ ์์
์ ์ค๋์ค๋ฅผ ํ
์คํธ๋ก ๋ณํํ๋ฏ๋ก ํน์ฑ ์ถ์ถ๊ธฐ์ ํ ํฌ๋์ด์ ๊ฐ ํ์ํฉ๋๋ค.
์ค๋์ค ์
๋ ฅ์ ์ฒ๋ฆฌํ ํน์ฑ ์ถ์ถ๊ธฐ๋ฅผ ๋ง๋ญ๋๋ค:
```py
>>> from transformers import Wav2Vec2FeatureExtractor
>>> feature_extractor = Wav2Vec2FeatureExtractor(padding_value=1.0, do_normalize=True)
```
ํ
์คํธ ์
๋ ฅ์ ์ฒ๋ฆฌํ ํ ํฌ๋์ด์ ๋ฅผ ๋ง๋ญ๋๋ค:
```py
>>> from transformers import Wav2Vec2CTCTokenizer
>>> tokenizer = Wav2Vec2CTCTokenizer(vocab_file="my_vocab_file.txt")
```
[`Wav2Vec2Processor`]์์ ํน์ฑ ์ถ์ถ๊ธฐ์ ํ ํฌ๋์ด์ ๋ฅผ ๊ฒฐํฉํฉ๋๋ค:
```py
>>> from transformers import Wav2Vec2Processor
>>> processor = Wav2Vec2Processor(feature_extractor=feature_extractor, tokenizer=tokenizer)
```
configuration๊ณผ ๋ชจ๋ธ์ด๋ผ๋ ๋ ๊ฐ์ง ๊ธฐ๋ณธ ํด๋์ค์ ์ถ๊ฐ ์ ์ฒ๋ฆฌ ํด๋์ค(ํ ํฌ๋์ด์ , ์ด๋ฏธ์ง ํ๋ก์ธ์, ํน์ฑ ์ถ์ถ๊ธฐ ๋๋ ํ๋ก์ธ์)๋ฅผ ์ฌ์ฉํ๋ฉด ๐ค Transformers์์ ์ง์ํ๋ ๋ชจ๋ ๋ชจ๋ธ์ ๋ง๋ค ์ ์์ต๋๋ค. ์ด๋ฌํ ๊ฐ ๊ธฐ๋ณธ ํด๋์ค๋ ๊ตฌ์ฑ์ด ๊ฐ๋ฅํ๋ฏ๋ก ์ํ๋ ํน์ ์์ฑ์ ์ฌ์ฉํ ์ ์์ต๋๋ค. ํ์ต์ ์ํด ๋ชจ๋ธ์ ์ฝ๊ฒ ์ค์ ํ๊ฑฐ๋ ๊ธฐ์กด์ ์ฌ์ ํ์ต๋ ๋ชจ๋ธ์ ์์ ํ์ฌ ๋ฏธ์ธ ์กฐ์ ํ ์ ์์ต๋๋ค.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/tf_xla.md
|
<!--Copyright 2023 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# TensorFlow ๋ชจ๋ธ์ ์ํ XLA ํตํฉ [[xla-integration-for-tensorflow-models]]
[[open-in-colab]]
XLA(Accelerated Linear Algebra)๋ TensorFlow ๋ชจ๋ธ์ ์คํ ์๊ฐ์ ๊ฐ์ํํ๊ธฐ ์ํ ์ปดํ์ผ๋ฌ์
๋๋ค. [๊ณต์ ๋ฌธ์](https://www.tensorflow.org/xla)์ ๋ฐ๋ฅด๋ฉด ๋ค์๊ณผ ๊ฐ์ต๋๋ค:
XLA(Accelerated Linear Algebra)๋ ์ ํ ๋์๋ฅผ ์ํ ๋๋ฉ์ธ ํนํ ์ปดํ์ผ๋ฌ๋ก, TensorFlow ๋ชจ๋ธ์ ์์ค ์ฝ๋ ๋ณ๊ฒฝ ์์ด ๊ฐ์ํํ ์ ์์ต๋๋ค.
TensorFlow์์ XLA๋ฅผ ์ฌ์ฉํ๋ ๊ฒ์ ๊ฐ๋จํฉ๋๋ค. XLA๋ `tensorflow` ๋ผ์ด๋ธ๋ฌ๋ฆฌ ๋ด์ ํจํค์ง๋ก ์ ๊ณต๋๋ฉฐ, [`tf.function`](https://www.tensorflow.org/guide/intro_to_graphs)๊ณผ ๊ฐ์ ๊ทธ๋ํ ์์ฑ ํจ์์์ `jit_compile` ์ธ์๋ฅผ ์ฌ์ฉํ์ฌ ํ์ฑํํ ์ ์์ต๋๋ค. `fit()` ๋ฐ `predict()`์ ๊ฐ์ Keras ๋ฉ์๋๋ฅผ ์ฌ์ฉํ๋ ๊ฒฝ์ฐ, `jit_compile` ์ธ์๋ฅผ `model.compile()`์ ์ ๋ฌํ์ฌ XLA๋ฅผ ๊ฐ๋จํ๊ฒ ํ์ฑํํ ์ ์์ต๋๋ค. ๊ทธ๋ฌ๋ XLA๋ ์ด๋ฌํ ๋ฉ์๋์ ๊ตญํ๋์ง ์๊ณ ์์์ `tf.function`์ ๊ฐ์ํํ๋ ๋ฐ์๋ ์ฌ์ฉํ ์ ์์ต๋๋ค.
๐ค Transformers์์๋ [GPT2](https://huggingface.co/docs/transformers/model_doc/gpt2), [T5](https://huggingface.co/docs/transformers/model_doc/t5), [OPT](https://huggingface.co/docs/transformers/model_doc/opt)์ ๊ฐ์ ๋ชจ๋ธ์ ํ
์คํธ ์์ฑ, ๊ทธ๋ฆฌ๊ณ [Whisper](https://huggingface.co/docs/transformers/model_doc/whisper)์ ๊ฐ์ ๋ชจ๋ธ์ ์์ฑ ์ฒ๋ฆฌ๋ฅผ ํฌํจํ์ฌ ์ฌ๋ฌ TensorFlow ๋ฉ์๋๊ฐ XLA์ ํธํ๋๋๋ก ๋ค์ ์์ฑ๋์์ต๋๋ค.
์ ํํ ์๋ ํฅ์์ ๋ชจ๋ธ์ ๋ฐ๋ผ ๋ค๋ฅด์ง๋ง, ๐ค Transformers ๋ด์ TensorFlow ํ
์คํธ ์์ฑ ๋ชจ๋ธ์ ๊ฒฝ์ฐ ์ต๋ 100๋ฐฐ์ ์๋ ํฅ์์ ํ์ธํ์ต๋๋ค. ์ด ๋ฌธ์์์๋ ์ด๋ฌํ ๋ชจ๋ธ์ ๋ํด XLA๋ฅผ ์ฌ์ฉํ์ฌ ์ต๋ ์ฑ๋ฅ์ ์ป๋ ๋ฐฉ๋ฒ์ ์ค๋ช
ํฉ๋๋ค. ๋ํ XLA ํตํฉ์ ๋ฒค์น๋งํฌ ๋ฐ ๋์์ธ ์ฒ ํ์ ๋ํ ์ถ๊ฐ ์๋ฃ ๋งํฌ๋ ์ ๊ณตํ ๊ฒ์
๋๋ค.
## XLA๋ฅผ ์ฌ์ฉํ์ฌ TF ํจ์ ์คํํ๊ธฐ [[running-tf-functions-with-xla]]
TensorFlow์์ ๋ค์๊ณผ ๊ฐ์ ๋ชจ๋ธ์ ๊ณ ๋ คํด ๋ด
์๋ค:
```py
import tensorflow as tf
model = tf.keras.Sequential(
[tf.keras.layers.Dense(10, input_shape=(10,), activation="relu"), tf.keras.layers.Dense(5, activation="softmax")]
)
```
์ ๋ชจ๋ธ์ ์ฐจ์์ด `(10, )`์ธ ์
๋ ฅ์ ๋ฐ์ต๋๋ค. ๋ค์๊ณผ ๊ฐ์ด ๋ชจ๋ธ์ ์ฌ์ฉํ์ฌ ์์ ํ๋ฅผ ์คํํ ์ ์์ต๋๋ค:
```py
# ๋ชจ๋ธ์ ๋ํ ์์์ ์
๋ ฅ์ ์์ฑํฉ๋๋ค.
batch_size = 16
input_vector_dim = 10
random_inputs = tf.random.normal((batch_size, input_vector_dim))
# ์์ ํ๋ฅผ ์คํํฉ๋๋ค.
_ = model(random_inputs)
```
XLA๋ก ์ปดํ์ผ๋ ํจ์๋ก ์์ ํ๋ฅผ ์คํํ๋ ค๋ฉด ๋ค์๊ณผ ๊ฐ์ด ํด์ผ ํฉ๋๋ค:
```py
xla_fn = tf.function(model, jit_compile=True)
_ = xla_fn(random_inputs)
```
`model`์ ๊ธฐ๋ณธ `call()` ํจ์๋ XLA ๊ทธ๋ํ๋ฅผ ์ปดํ์ผํ๋ ๋ฐ ์ฌ์ฉ๋ฉ๋๋ค. ๊ทธ๋ฌ๋ ๋ค๋ฅธ ๋ชจ๋ธ ํจ์๋ฅผ XLA๋ก ์ปดํ์ผํ๋ ค๋ฉด ๋ค์๊ณผ ๊ฐ์ด ํ ์๋ ์์ต๋๋ค:
```py
my_xla_fn = tf.function(model.my_xla_fn, jit_compile=True)
```
## ๐ค Transformers์์ XLA๋ฅผ ์ฌ์ฉํ์ฌ TF ํ
์คํธ ์์ฑ ๋ชจ๋ธ ์คํํ๊ธฐ [[running-a-tf-text-generation-model-with-xla-from-transformers]]
๐ค Transformers์์ XLA๋ก ๊ฐ์ํ๋ ์์ฑ์ ํ์ฑํํ๋ ค๋ฉด ์ต์ ๋ฒ์ ์ `transformers`๊ฐ ์ค์น๋์ด ์์ด์ผ ํฉ๋๋ค. ๋ค์๊ณผ ๊ฐ์ด ์ค์นํ ์ ์์ต๋๋ค:
```bash
pip install transformers --upgrade
```
๊ทธ๋ฆฌ๊ณ ๋ค์ ์ฝ๋๋ฅผ ์คํํ ์ ์์ต๋๋ค:
```py
import tensorflow as tf
from transformers import AutoTokenizer, TFAutoModelForCausalLM
# ์ต์ ๋ฒ์ ์ Transformers๊ฐ ์ค์น๋์ด ์์ง ์๋ค๋ฉด ์ค๋ฅ๊ฐ ๋ฐ์ํฉ๋๋ค.
from transformers.utils import check_min_version
check_min_version("4.21.0")
tokenizer = AutoTokenizer.from_pretrained("gpt2", padding_side="left", pad_token="</s>")
model = TFAutoModelForCausalLM.from_pretrained("gpt2")
input_string = ["TensorFlow is"]
# XLA ์์ฑ ํจ์๋ฅผ ๋ง๋ค๊ธฐ ์ํ ํ ์ค
xla_generate = tf.function(model.generate, jit_compile=True)
tokenized_input = tokenizer(input_string, return_tensors="tf")
generated_tokens = xla_generate(**tokenized_input, num_beams=2)
decoded_text = tokenizer.decode(generated_tokens[0], skip_special_tokens=True)
print(f"Generated -- {decoded_text}")
# Generated -- TensorFlow is an open-source, open-source, distributed-source application # framework for the
```
์ ์ ์๋ฏ์ด, `generate()`์์ XLA๋ฅผ ํ์ฑํํ๋ ๊ฒ์ ๋จ ํ ์ค์ ์ฝ๋์
๋๋ค. ์ฝ๋์ ๋๋จธ์ง ๋ถ๋ถ์ ๋ณ๊ฒฝ๋์ง ์์ต๋๋ค. ๊ทธ๋ฌ๋ ์ ์ฝ๋ ์ค๋ํซ์์๋ XLA์ ํน์ ํ ๋ช ๊ฐ์ง ์ฃผ์ํ ์ ์ด ์์ต๋๋ค. XLA๊ฐ ๊ฐ์ ธ๋ค์ค ์๋ ํฅ์์ ์คํํ๊ธฐ ์ํด์๋ ์ด๋ฅผ ์๊ณ ์์ด์ผ ํฉ๋๋ค. ๋ค์ ์น์
์์ ์ด์ ๋ํด ๋
ผ์ํฉ๋๋ค.
## ์ฃผ์ํ ์ [[gotchas-to-be-aware-of]]
XLA ํ์ฑํ ํจ์(`xla_generate()`์ ๊ฐ์)๋ฅผ ์ฒ์ ์คํํ ๋ ๋ด๋ถ์ ์ผ๋ก ๊ณ์ฐ ๊ทธ๋ํ๋ฅผ ์ถ๋ก ํ๋ ค๊ณ ํ๋ฉฐ, ์ด๋ ์๊ฐ์ด ์์๋ฉ๋๋ค. ์ด ๊ณผ์ ์ [โ์ถ์ (tracing)โ](https://www.tensorflow.org/guide/intro_to_graphs#when_is_a_function_tracing)์ด๋ผ๊ณ ์๋ ค์ ธ ์์ต๋๋ค.
์์ฑ ์๊ฐ์ด ๋น ๋ฅด์ง ์๋ค๋ ๊ฒ์ ์ ์ ์์ ๊ฒ์
๋๋ค. `xla_generate()`(๋๋ ๋ค๋ฅธ XLA ํ์ฑํ ํจ์)์ ์ฐ์ ํธ์ถ์ ํจ์์ ์ ๋ฌ๋ ์
๋ ฅ์ด ์ด๊ธฐ์ ๊ตฌ์ถ๋ ๊ณ์ฐ ๊ทธ๋ํ์ ๋์ผํ ํํ๋ฅผ ๋ฐ๋ฅธ๋ค๋ฉด, ๊ณ์ฐ ๊ทธ๋ํ๋ฅผ ์ถ๋ก ํ ํ์๊ฐ ์์ต๋๋ค. ์ด๋ ์
๋ ฅ ํํ๊ฐ ๊ณ ์ ๋ ๋ชจ๋ฌ๋ฆฌํฐ(์: ์ด๋ฏธ์ง)์๋ ๋ฌธ์ ๊ฐ ๋์ง ์์ง๋ง, ๊ฐ๋ณ ์
๋ ฅ ํํ ๋ชจ๋ฌ๋ฆฌํฐ(์: ํ
์คํธ)๋ฅผ ์ฌ์ฉํ ๋ ์ฃผ์ํด์ผ ํฉ๋๋ค.
`xla_generate()`๊ฐ ํญ์ ๋์ผํ ์
๋ ฅ ํํ๋ก ๋์ํ๋๋ก ํ๋ ค๋ฉด, ํ ํฌ๋์ด์ ๋ฅผ ํธ์ถํ ๋ `padding` ์ธ์๋ฅผ ์ง์ ํ ์ ์์ต๋๋ค.
```py
import tensorflow as tf
from transformers import AutoTokenizer, TFAutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("gpt2", padding_side="left", pad_token="</s>")
model = TFAutoModelForCausalLM.from_pretrained("gpt2")
input_string = ["TensorFlow is"]
xla_generate = tf.function(model.generate, jit_compile=True)
# ์ฌ๊ธฐ์, padding ์ต์
์ด ์๋ ํ ํฌ๋์ด์ ๋ฅผ ํธ์ถํฉ๋๋ค.
tokenized_input = tokenizer(input_string, pad_to_multiple_of=8, padding=True, return_tensors="tf")
generated_tokens = xla_generate(**tokenized_input, num_beams=2)
decoded_text = tokenizer.decode(generated_tokens[0], skip_special_tokens=True)
print(f"Generated -- {decoded_text}")
```
์ด๋ ๊ฒ ํ๋ฉด `xla_generate()`์ ๋ํ ์
๋ ฅ์ด ํญ์ ์ถ์ ๋ ํํ๋ก ์ ๋ฌ๋์ด ์์ฑ ์๊ฐ์ด ๊ฐ์ํ๋ฉ๋๋ค. ๋ค์ ์ฝ๋๋ก ์ด๋ฅผ ํ์ธํ ์ ์์ต๋๋ค:
```py
import time
import tensorflow as tf
from transformers import AutoTokenizer, TFAutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("gpt2", padding_side="left", pad_token="</s>")
model = TFAutoModelForCausalLM.from_pretrained("gpt2")
xla_generate = tf.function(model.generate, jit_compile=True)
for input_string in ["TensorFlow is", "TensorFlow is a", "TFLite is a"]:
tokenized_input = tokenizer(input_string, pad_to_multiple_of=8, padding=True, return_tensors="tf")
start = time.time_ns()
generated_tokens = xla_generate(**tokenized_input, num_beams=2)
end = time.time_ns()
print(f"Execution time -- {(end - start) / 1e6:.1f} ms\n")
```
Tesla T4 GPU์์๋ ๋ค์๊ณผ ๊ฐ์ ์ถ๋ ฅ์ ์์ํ ์ ์์ต๋๋ค:
```bash
Execution time -- 30819.6 ms
Execution time -- 79.0 ms
Execution time -- 78.9 ms
```
`xla_generate()`์ ์ฒซ ๋ฒ์งธ ํธ์ถ์ ์ถ์ ๋๋ฌธ์ ์๊ฐ์ด ์ค๋ ๊ฑธ๋ฆฌ์ง๋ง, ์ฐ์ ํธ์ถ์ ๋ช ๋ฐฐ๋ ๋น ๋ฆ
๋๋ค. ์์ฑ ์ต์
์ ๋ํ ์ด๋ค ๋ณ๊ฒฝ์ด๋ ๋ค์ ์ถ์ ์ ์ ๋ฐํ๋ฏ๋ก ์์ฑ ์๊ฐ์ด ๋๋ ค์ง ์ ์์์ ๋ช
์ฌํ์ธ์.
์ด ๋ฌธ์์์๋ ๐ค Transformers์์ ์ ๊ณตํ๋ ๋ชจ๋ ํ
์คํธ ์์ฑ ์ต์
์ ๋ค๋ฃจ์ง ์์์ต๋๋ค. ๊ณ ๊ธ ์ฌ์ฉ ์ฌ๋ก์ ๋ํด ๋ฌธ์๋ฅผ ์ฐธ์กฐํ์๊ธฐ ๋ฐ๋๋๋ค.
## ์ถ๊ฐ ์๋ฃ [[additional-resources]]
์ฌ๊ธฐ์ ๐ค Transformers์ XLA์ ๋ํด ๋ ์์ธํ ์๊ณ ์ถ์ ๊ฒฝ์ฐ ๋์์ด ๋ ์ ์๋ ๋ช ๊ฐ์ง ์ถ๊ฐ ์๋ฃ๋ฅผ ์ ๊ณตํฉ๋๋ค.
* [์ด Colab ๋
ธํธ๋ถ](https://colab.research.google.com/github/huggingface/blog/blob/main/notebooks/91_tf_xla_generate.ipynb)์ XLA์ ํธํ๋๋ ์ธ์ฝ๋-๋์ฝ๋([T5](https://huggingface.co/docs/transformers/model_doc/t5)์ ๊ฐ์) ๋ฐ ๋์ฝ๋ ์ ์ฉ([GPT2](https://huggingface.co/docs/transformers/model_doc/gpt2)์ ๊ฐ์) ํ
์คํธ ์์ฑ ๋ชจ๋ธ์ ์คํํด ๋ณผ ์ ์๋ ๋ํํ ๋ฐ๋ชจ๋ฅผ ์ ๊ณตํฉ๋๋ค.
* [์ด ๋ธ๋ก๊ทธ ๊ธ](https://huggingface.co/blog/tf-xla-generate)์ TensorFlow์์ XLA์ ๋ํ ์น์ ํ ์๊ฐ์ ํจ๊ป XLA์ ํธํ๋๋ ๋ชจ๋ธ์ ๋น๊ต ๋ฒค์น๋งํฌ์ ๋ํ ๊ฐ์๋ฅผ ์ ๊ณตํฉ๋๋ค.
* [์ด ๋ธ๋ก๊ทธ ๊ธ](https://blog.tensorflow.org/2022/11/how-hugging-face-improved-text-generation-performance-with-xla.html)์ ๐ค Transformers์ TensorFlow ๋ชจ๋ธ์ XLA ์ง์์ ์ถ๊ฐํ๋ ๊ฒ์ ๋ํ ๋์์ธ ์ฒ ํ์ ๋
ผ์ํฉ๋๋ค.
* XLA์ TensorFlow ๊ทธ๋ํ์ ๋ํด ๋ ์์ธํ ์๊ณ ์ถ์ ๊ฒฝ์ฐ ์ถ์ฒํ๋ ๊ธ:
* [XLA: ๊ธฐ๊ณ ํ์ต์ ์ํ ์ต์ ํ ์ปดํ์ผ๋ฌ](https://www.tensorflow.org/xla)
* [๊ทธ๋ํ ๋ฐ tf.function ์๊ฐ](https://www.tensorflow.org/guide/intro_to_graphs)
* [tf.function์ผ๋ก ์ฑ๋ฅ ํฅ์ํ๊ธฐ](https://www.tensorflow.org/guide/function)
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/installation.md
|
<!---
Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ์ค์น๋ฐฉ๋ฒ[[installation]]
๐ค Transformers๋ฅผ ์ฌ์ฉ ์ค์ธ ๋ฅ๋ฌ๋ ๋ผ์ด๋ธ๋ฌ๋ฆฌ์ ๋ง์ถฐ ์ค์นํ๊ณ , ์บ์๋ฅผ ๊ตฌ์ฑํ๊ฑฐ๋ ์ ํ์ ์ผ๋ก ์คํ๋ผ์ธ์์๋ ์คํํ ์ ์๋๋ก ๐ค Transformers๋ฅผ ์ค์ ํ๋ ๋ฐฉ๋ฒ์ ๋ฐฐ์ฐ๊ฒ ์ต๋๋ค.
๐ค Transformers๋ Python 3.6+, PyTorch 1.1.0+, TensorFlow 2.0+ ๋ฐ Flax์์ ํ
์คํธ๋์์ต๋๋ค. ๋ฅ๋ฌ๋ ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ์ค์นํ๋ ค๋ฉด ์๋ ๋งํฌ๋ ์ ๋ง๋ค์ ๊ณต์ ์ฌ์ดํธ๋ฅผ ์ฐธ๊ณ ํด์ฃผ์ธ์.
* [PyTorch](https://pytorch.org/get-started/locally/) ์ค์นํ๊ธฐ
* [TensorFlow 2.0](https://www.tensorflow.org/install/pip) ์ค์นํ๊ธฐ
* [Flax](https://flax.readthedocs.io/en/latest/) ์ค์นํ๊ธฐ
## pip์ผ๋ก ์ค์นํ๊ธฐ[[install-with-pip]]
๐ค Transformers๋ฅผ [๊ฐ์ ํ๊ฒฝ](https://docs.python.org/3/library/venv.html)์ ์ค์นํ๋ ๊ฒ์ ์ถ์ฒ๋๋ฆฝ๋๋ค. Python ๊ฐ์ ํ๊ฒฝ์ ์ต์ํ์ง ์๋ค๋ฉด, ์ด [๊ฐ์ด๋](https://packaging.python.org/guides/installing-using-pip-and-virtual-environments/)๋ฅผ ์ฐธ๊ณ ํ์ธ์. ๊ฐ์ ํ๊ฒฝ์ ์ฌ์ฉํ๋ฉด ์๋ก ๋ค๋ฅธ ํ๋ก์ ํธ๋ค์ ๋ณด๋ค ์ฝ๊ฒ ๊ด๋ฆฌํ ์ ์๊ณ , ์์กด์ฑ ๊ฐ์ ํธํ์ฑ ๋ฌธ์ ๋ฅผ ๋ฐฉ์งํ ์ ์์ต๋๋ค.
๋จผ์ ํ๋ก์ ํธ ๋๋ ํ ๋ฆฌ์์ ๊ฐ์ ํ๊ฒฝ์ ๋ง๋ค์ด ์ค๋๋ค.
```bash
python -m venv .env
```
๊ฐ์ ํ๊ฒฝ์ ํ์ฑํํด์ฃผ์ธ์. Linux๋ MacOS์ ๊ฒฝ์ฐ:
```bash
source .env/bin/activate
```
Windows์ ๊ฒฝ์ฐ:
```bash
.env/Scripts/activate
```
์ด์ ๐ค Transformers๋ฅผ ์ค์นํ ์ค๋น๊ฐ ๋์์ต๋๋ค. ๋ค์ ๋ช
๋ น์ ์
๋ ฅํด์ฃผ์ธ์.
```bash
pip install transformers
```
CPU๋ง ์จ๋ ๋๋ค๋ฉด, ๐ค Transformers์ ๋ฅ๋ฌ๋ ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ๋จ 1์ค๋ก ์ค์นํ ์ ์์ต๋๋ค. ์๋ฅผ ๋ค์ด ๐ค Transformers์ PyTorch์ ๊ฒฝ์ฐ:
```bash
pip install transformers[torch]
```
๐ค Transformers์ TensorFlow 2.0์ ๊ฒฝ์ฐ:
```bash
pip install transformers[tf-cpu]
```
๐ค Transformers์ Flax์ ๊ฒฝ์ฐ:
```bash
pip install transformers[flax]
```
๋ง์ง๋ง์ผ๋ก ๐ค Transformers๊ฐ ์ ๋๋ก ์ค์น๋์๋์ง ํ์ธํ ์ฐจ๋ก์
๋๋ค. ์ฌ์ ํ๋ จ๋ ๋ชจ๋ธ์ ๋ค์ด๋ก๋ํ๋ ์ฝ๋์
๋๋ค.
```bash
python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('we love you'))"
```
๋ผ๋ฒจ๊ณผ ์ ์๊ฐ ์ถ๋ ฅ๋๋ฉด ์ ์ค์น๋ ๊ฒ์
๋๋ค.
```bash
[{'label': 'POSITIVE', 'score': 0.9998704791069031}]
```
## ์์ค์์ ์ค์นํ๊ธฐ[[install-from-source]]
๐ค Transformers๋ฅผ ์์ค์์ ์ค์นํ๋ ค๋ฉด ์๋ ๋ช
๋ น์ ์คํํ์ธ์.
```bash
pip install git+https://github.com/huggingface/transformers
```
์ ๋ช
๋ น์ ์ต์ ์ด์ง๋ง (์์ ์ ์ธ) `stable` ๋ฒ์ ์ด ์๋ ์คํ์ฑ์ด ์ง์ `main` ๋ฒ์ ์ ์ค์นํฉ๋๋ค. `main` ๋ฒ์ ์ ๊ฐ๋ฐ ํํฉ๊ณผ ๋ฐ๋ง์ถ๋๋ฐ ์ ์ฉํฉ๋๋ค. ์์๋ก ๋ง์ง๋ง ๊ณต์ ๋ฆด๋ฆฌ์ค ์ดํ ๋ฐ๊ฒฌ๋ ๋ฒ๊ทธ๊ฐ ํจ์น๋์์ง๋ง, ์ ๋ฆด๋ฆฌ์ค๋ก ์์ง ๋กค์์๋์ง๋ ์์ ๊ฒฝ์ฐ๋ฅผ ๋ค ์ ์์ต๋๋ค. ๋ฐ๊ฟ ๋งํ๋ฉด `main` ๋ฒ์ ์ด ์์ ์ฑ๊ณผ๋ ๊ฑฐ๋ฆฌ๊ฐ ์๋ค๋ ๋ป์ด๊ธฐ๋ ํฉ๋๋ค. ์ ํฌ๋ `main` ๋ฒ์ ์ ์ฌ์ฉํ๋๋ฐ ๋ฌธ์ ๊ฐ ์๋๋ก ๋
ธ๋ ฅํ๊ณ ์์ผ๋ฉฐ, ๋๋ถ๋ถ์ ๋ฌธ์ ๋ ๋๊ฐ ๋ช ์๊ฐ์ด๋ ํ๋ฃจ ์์ ํด๊ฒฐ๋ฉ๋๋ค. ๋ง์ฝ ๋ฌธ์ ๊ฐ ๋ฐ์ํ๋ฉด [์ด์](https://github.com/huggingface/transformers/issues)๋ฅผ ์ด์ด์ฃผ์๋ฉด ๋ ๋นจ๋ฆฌ ํด๊ฒฐํ ์ ์์ต๋๋ค!
์ ๊ณผ ๋ง์ฐฌ๊ฐ์ง๋ก ๐ค Transformers๊ฐ ์ ๋๋ก ์ค์น๋์๋์ง ํ์ธํ ์ฐจ๋ก์
๋๋ค.
```bash
python -c "from transformers import pipeline; print(pipeline('sentiment-analysis')('I love you'))"
```
## ์์ ๊ฐ๋ฅํ ์ค์น[[editable-install]]
์์ ๊ฐ๋ฅํ ์ค์น๊ฐ ํ์ํ ๊ฒฝ์ฐ๋ ๋ค์๊ณผ ๊ฐ์ต๋๋ค.
* `main` ๋ฒ์ ์ ์์ค ์ฝ๋๋ฅผ ์ฌ์ฉํ๊ธฐ ์ํด
* ๐ค Transformers์ ๊ธฐ์ฌํ๊ณ ์ถ์ด์ ์ฝ๋์ ๋ณ๊ฒฝ ์ฌํญ์ ํ
์คํธํ๊ธฐ ์ํด
๋ฆฌํฌ์งํฐ๋ฆฌ๋ฅผ ๋ณต์ ํ๊ณ ๐ค Transformers๋ฅผ ์ค์นํ๋ ค๋ฉด ๋ค์ ๋ช
๋ น์ ์
๋ ฅํด์ฃผ์ธ์.
```bash
git clone https://github.com/huggingface/transformers.git
cd transformers
pip install -e .
```
์ ๋ช
๋ น์ ๋ฆฌํฌ์งํฐ๋ฆฌ๋ฅผ ๋ณต์ ํ ์์น์ ํด๋์ Python ๋ผ์ด๋ธ๋ฌ๋ฆฌ์ ๊ฒฝ๋ก๋ฅผ ์ฐ๊ฒฐ์ํต๋๋ค. Python์ด ์ผ๋ฐ ๋ผ์ด๋ธ๋ฌ๋ฆฌ ๊ฒฝ๋ก ์ธ์ ๋ณต์ ํ ํด๋ ๋ด๋ถ๋ฅผ ํ์ธํ ๊ฒ์
๋๋ค. ์๋ฅผ ๋ค์ด Python ํจํค์ง๊ฐ ์ผ๋ฐ์ ์ผ๋ก `~/anaconda3/envs/main/lib/python3.7/site-packages/`์ ์ค์น๋์ด ์๋๋ฐ, ๋ช
๋ น์ ๋ฐ์ Python์ด ์ด์ ๋ณต์ ํ ํด๋์ธ `~/transformers/`๋ ๊ฒ์ํ๊ฒ ๋ฉ๋๋ค.
<Tip warning={true}>
๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ๊ณ์ ์ฌ์ฉํ๋ ค๋ฉด `transformers` ํด๋๋ฅผ ๊ผญ ์ ์งํด์ผ ํฉ๋๋ค.
</Tip>
๋ณต์ ๋ณธ์ ์ต์ ๋ฒ์ ์ ๐ค Transformers๋ก ์ฝ๊ฒ ์
๋ฐ์ดํธํ ์ ์์ต๋๋ค.
```bash
cd ~/transformers/
git pull
```
Python ํ๊ฒฝ์ ๋ค์ ์คํํ๋ฉด ์
๋ฐ์ดํธ๋ ๐ค Transformers์ `main` ๋ฒ์ ์ ์ฐพ์๋ผ ๊ฒ์
๋๋ค.
## conda๋ก ์ค์นํ๊ธฐ[[install-with-conda]]
`huggingface` conda ์ฑ๋์์ ์ค์นํ ์ ์์ต๋๋ค.
```bash
conda install -c huggingface transformers
```
## ์บ์ ๊ตฌ์ฑํ๊ธฐ[[cache-setup]]
์ฌ์ ํ๋ จ๋ ๋ชจ๋ธ์ ๋ค์ด๋ก๋๋ ํ ๋ก์ปฌ ๊ฒฝ๋ก `~/.cache/huggingface/hub`์ ์บ์๋ฉ๋๋ค. ์
ธ ํ๊ฒฝ ๋ณ์ `TRANSFORMERS_CACHE`์ ๊ธฐ๋ณธ ๋๋ ํฐ๋ฆฌ์
๋๋ค. Windows์ ๊ฒฝ์ฐ ๊ธฐ๋ณธ ๋๋ ํฐ๋ฆฌ๋ `C:\Users\username\.cache\huggingface\hub`์
๋๋ค. ์๋์ ์
ธ ํ๊ฒฝ ๋ณ์๋ฅผ (์ฐ์ ์์) ์์๋๋ก ๋ณ๊ฒฝํ์ฌ ๋ค๋ฅธ ์บ์ ๋๋ ํ ๋ฆฌ๋ฅผ ์ง์ ํ ์ ์์ต๋๋ค.
1. ์
ธ ํ๊ฒฝ ๋ณ์ (๊ธฐ๋ณธ): `HUGGINGFACE_HUB_CACHE` ๋๋ `TRANSFORMERS_CACHE`
2. ์
ธ ํ๊ฒฝ ๋ณ์: `HF_HOME`
3. ์
ธ ํ๊ฒฝ ๋ณ์: `XDG_CACHE_HOME` + `/huggingface`
<Tip>
๊ณผ๊ฑฐ ๐ค Transformers์์ ์ฐ์๋ ์
ธ ํ๊ฒฝ ๋ณ์ `PYTORCH_TRANSFORMERS_CACHE` ๋๋ `PYTORCH_PRETRAINED_BERT_CACHE`์ด ์ค์ ๋์๋ค๋ฉด, ์
ธ ํ๊ฒฝ ๋ณ์ `TRANSFORMERS_CACHE`์ ์ง์ ํ์ง ์๋ ํ ์ฐ์ ์ฌ์ฉ๋ฉ๋๋ค.
</Tip>
## ์คํ๋ผ์ธ ๋ชจ๋[[offline-mode]]
๐ค Transformers๋ฅผ ๋ก์ปฌ ํ์ผ๋ง ์ฌ์ฉํ๋๋ก ํด์ ๋ฐฉํ๋ฒฝ ๋๋ ์คํ๋ผ์ธ ํ๊ฒฝ์์ ์คํํ ์ ์์ต๋๋ค. ํ์ฑํํ๋ ค๋ฉด `TRANSFORMERS_OFFLINE=1` ํ๊ฒฝ ๋ณ์๋ฅผ ์ค์ ํ์ธ์.
<Tip>
`HF_DATASETS_OFFLINE=1` ํ๊ฒฝ ๋ณ์๋ฅผ ์ค์ ํ์ฌ ์คํ๋ผ์ธ ํ๋ จ ๊ณผ์ ์ [๐ค Datasets](https://huggingface.co/docs/datasets/)์ ์ถ๊ฐํ ์ ์์ต๋๋ค.
</Tip>
์๋ฅผ ๋ค์ด ์ธ๋ถ ๊ธฐ๊ธฐ ์ฌ์ด์ ๋ฐฉํ๋ฒฝ์ ๋ ์ผ๋ฐ ๋คํธ์ํฌ์์ ํ์์ฒ๋ผ ํ๋ก๊ทธ๋จ์ ๋ค์๊ณผ ๊ฐ์ด ์คํํ ์ ์์ต๋๋ค.
```bash
python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
```
์คํ๋ผ์ธ ๊ธฐ๊ธฐ์์ ๋์ผํ ํ๋ก๊ทธ๋จ์ ๋ค์๊ณผ ๊ฐ์ด ์คํํ ์ ์์ต๋๋ค.
```bash
HF_DATASETS_OFFLINE=1 TRANSFORMERS_OFFLINE=1 \
python examples/pytorch/translation/run_translation.py --model_name_or_path t5-small --dataset_name wmt16 --dataset_config ro-en ...
```
์ด์ ์คํฌ๋ฆฝํธ๋ ๋ก์ปฌ ํ์ผ์ ํํด์๋ง ๊ฒ์ํ ๊ฒ์ด๋ฏ๋ก, ์คํฌ๋ฆฝํธ๊ฐ ์ค๋จ๋๊ฑฐ๋ ์๊ฐ์ด ์ด๊ณผ๋ ๋๊น์ง ๋ฉ์ถฐ์์ง ์๊ณ ์ ์คํ๋ ๊ฒ์
๋๋ค.
### ์คํ๋ผ์ธ์ฉ ๋ชจ๋ธ ๋ฐ ํ ํฌ๋์ด์ ๋ง๋ค์ด๋๊ธฐ[[fetch-models-and-tokenizers-to-use-offline]]
Another option for using ๐ค Transformers offline is to download the files ahead of time, and then point to their local path when you need to use them offline. There are three ways to do this:
๐ค Transformers๋ฅผ ์คํ๋ผ์ธ์ผ๋ก ์ฌ์ฉํ๋ ๋ ๋ค๋ฅธ ๋ฐฉ๋ฒ์ ํ์ผ์ ๋ฏธ๋ฆฌ ๋ค์ด๋ก๋ํ ๋ค์, ์คํ๋ผ์ธ์ผ ๋ ์ฌ์ฉํ ๋ก์ปฌ ๊ฒฝ๋ก๋ฅผ ์ง์ ํด๋๋ ๊ฒ์
๋๋ค. 3๊ฐ์ง ์ค ํธํ ๋ฐฉ๋ฒ์ ๊ณ ๋ฅด์ธ์.
* [Model Hub](https://huggingface.co/models)์ UI๋ฅผ ํตํด ํ์ผ์ ๋ค์ด๋ก๋ํ๋ ค๋ฉด โ ์์ด์ฝ์ ํด๋ฆญํ์ธ์.
* [`PreTrainedModel.from_pretrained`]์ [`PreTrainedModel.save_pretrained`] ์ํฌํ๋ก๋ฅผ ํ์ฉํ์ธ์.
1. ๋ฏธ๋ฆฌ [`PreTrainedModel.from_pretrained`]๋ก ํ์ผ์ ๋ค์ด๋ก๋ํด๋์ธ์.
```py
>>> from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
>>> tokenizer = AutoTokenizer.from_pretrained("bigscience/T0_3B")
>>> model = AutoModelForSeq2SeqLM.from_pretrained("bigscience/T0_3B")
```
2. [`PreTrainedModel.save_pretrained`]๋ก ์ง์ ๋ ๊ฒฝ๋ก์ ํ์ผ์ ์ ์ฅํด๋์ธ์.
```py
>>> tokenizer.save_pretrained("./your/path/bigscience_t0")
>>> model.save_pretrained("./your/path/bigscience_t0")
```
3. ์ด์ ์คํ๋ผ์ธ์ผ ๋ [`PreTrainedModel.from_pretrained`]๋ก ์ ์ฅํด๋๋ ํ์ผ์ ์ง์ ๋ ๊ฒฝ๋ก์์ ๋ค์ ๋ถ๋ฌ์ค์ธ์.
```py
>>> tokenizer = AutoTokenizer.from_pretrained("./your/path/bigscience_t0")
>>> model = AutoModel.from_pretrained("./your/path/bigscience_t0")
```
* [huggingface_hub](https://github.com/huggingface/huggingface_hub/tree/main/src/huggingface_hub) ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ํ์ฉํด์ ํ์ผ์ ๋ค์ด๋ก๋ํ์ธ์.
1. ๊ฐ์ํ๊ฒฝ์ `huggingface_hub` ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ์ค์นํ์ธ์.
```bash
python -m pip install huggingface_hub
```
2. [`hf_hub_download`](https://huggingface.co/docs/hub/adding-a-library#download-files-from-the-hub) ํจ์๋ก ํ์ผ์ ํน์ ์์น์ ๋ค์ด๋ก๋ํ ์ ์์ต๋๋ค. ์๋ฅผ ๋ค์ด ์๋ ๋ช
๋ น์ [T0](https://huggingface.co/bigscience/T0_3B) ๋ชจ๋ธ์ `config.json` ํ์ผ์ ์ง์ ๋ ๊ฒฝ๋ก์ ๋ค์ด๋ก๋ํฉ๋๋ค.
```py
>>> from huggingface_hub import hf_hub_download
>>> hf_hub_download(repo_id="bigscience/T0_3B", filename="config.json", cache_dir="./your/path/bigscience_t0")
```
ํ์ผ์ ๋ค์ด๋ก๋ํ๊ณ ๋ก์ปฌ์ ์บ์ ํด๋๊ณ ๋๋ฉด, ๋์ค์ ๋ถ๋ฌ์ ์ฌ์ฉํ ์ ์๋๋ก ๋ก์ปฌ ๊ฒฝ๋ก๋ฅผ ์ง์ ํด๋์ธ์.
```py
>>> from transformers import AutoConfig
>>> config = AutoConfig.from_pretrained("./your/path/bigscience_t0/config.json")
```
<Tip>
Hub์ ์ ์ฅ๋ ํ์ผ์ ๋ค์ด๋ก๋ํ๋ ๋ฐฉ๋ฒ์ ๋ ์์ธํ ์์๋ณด๋ ค๋ฉด [Hub์์ ํ์ผ ๋ค์ด๋ก๋ํ๊ธฐ](https://huggingface.co/docs/hub/how-to-downstream) ์น์
์ ์ฐธ๊ณ ํด์ฃผ์ธ์.
</Tip>
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/hpo_train.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# Trainer API๋ฅผ ์ฌ์ฉํ ํ์ดํผํ๋ผ๋ฏธํฐ ํ์ [[hyperparameter-search-using-trainer-api]]
๐ค Transformers์์๋ ๐ค Transformers ๋ชจ๋ธ์ ํ์ต์ํค๋๋ฐ ์ต์ ํ๋ [`Trainer`] ํด๋์ค๋ฅผ ์ ๊ณตํ๊ธฐ ๋๋ฌธ์, ์ฌ์ฉ์๋ ์ง์ ํ๋ จ ๋ฃจํ๋ฅผ ์์ฑํ ํ์ ์์ด ๋์ฑ ๊ฐํธํ๊ฒ ํ์ต์ ์ํฌ ์ ์์ต๋๋ค. ๋ํ, [`Trainer`]๋ ํ์ดํผํ๋ผ๋ฏธํฐ ํ์์ ์ํ API๋ฅผ ์ ๊ณตํฉ๋๋ค. ์ด ๋ฌธ์์์ ์ด API๋ฅผ ํ์ฉํ๋ ๋ฐฉ๋ฒ์ ์์์ ํจ๊ป ๋ณด์ฌ๋๋ฆฌ๊ฒ ์ต๋๋ค.
## ํ์ดํผํ๋ผ๋ฏธํฐ ํ์ ๋ฐฑ์๋ [[hyperparameter-search-backend]]
[`Trainer`]๋ ํ์ฌ ์๋ 4๊ฐ์ง ํ์ดํผํ๋ผ๋ฏธํฐ ํ์ ๋ฐฑ์๋๋ฅผ ์ง์ํฉ๋๋ค:
[optuna](https://optuna.org/)์ [sigopt](https://sigopt.com/), [raytune](https://docs.ray.io/en/latest/tune/index.html), [wandb](https://wandb.ai/site/sweeps) ์
๋๋ค.
ํ์ดํผํ๋ผ๋ฏธํฐ ํ์ ๋ฐฑ์๋๋ก ์ฌ์ฉํ๊ธฐ ์ ์ ์๋์ ๋ช
๋ น์ด๋ฅผ ์ฌ์ฉํ์ฌ ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ค์ ์ค์นํ์ธ์.
```bash
pip install optuna/sigopt/wandb/ray[tune]
```
## ์์ ์์ ํ์ดํผํ๋ผ๋ฏธํฐ ํ์์ ํ์ฑํํ๋ ๋ฐฉ๋ฒ [[how-to-enable-hyperparameter-search-in-example]]
ํ์ดํผํ๋ผ๋ฏธํฐ ํ์ ๊ณต๊ฐ์ ์ ์ํ์ธ์. ํ์ดํผํ๋ผ๋ฏธํฐ ํ์ ๋ฐฑ์๋๋ง๋ค ์๋ก ๋ค๋ฅธ ํ์์ด ํ์ํฉ๋๋ค.
sigopt์ ๊ฒฝ์ฐ, ํด๋น [object_parameter](https://docs.sigopt.com/ai-module-api-references/api_reference/objects/object_parameter) ๋ฌธ์๋ฅผ ์ฐธ์กฐํ์ฌ ์๋์ ๊ฐ์ด ์์ฑํ์ธ์:
```py
>>> def sigopt_hp_space(trial):
... return [
... {"bounds": {"min": 1e-6, "max": 1e-4}, "name": "learning_rate", "type": "double"},
... {
... "categorical_values": ["16", "32", "64", "128"],
... "name": "per_device_train_batch_size",
... "type": "categorical",
... },
... ]
```
optuna์ ๊ฒฝ์ฐ, ํด๋น [object_parameter](https://optuna.readthedocs.io/en/stable/tutorial/10_key_features/002_configurations.html#sphx-glr-tutorial-10-key-features-002-configurations-py) ๋ฌธ์๋ฅผ ์ฐธ์กฐํ์ฌ ์๋์ ๊ฐ์ด ์์ฑํ์ธ์:
```py
>>> def optuna_hp_space(trial):
... return {
... "learning_rate": trial.suggest_float("learning_rate", 1e-6, 1e-4, log=True),
... "per_device_train_batch_size": trial.suggest_categorical("per_device_train_batch_size", [16, 32, 64, 128]),
... }
```
raytune์ ๊ฒฝ์ฐ, ํด๋น [object_parameter](https://docs.ray.io/en/latest/tune/api/search_space.html) ๋ฌธ์๋ฅผ ์ฐธ์กฐํ์ฌ ์๋์ ๊ฐ์ด ์์ฑํ์ธ์:
```py
>>> def ray_hp_space(trial):
... return {
... "learning_rate": tune.loguniform(1e-6, 1e-4),
... "per_device_train_batch_size": tune.choice([16, 32, 64, 128]),
... }
```
wandb์ ๊ฒฝ์ฐ, ํด๋น [object_parameter](https://docs.wandb.ai/guides/sweeps/configuration) ๋ฌธ์๋ฅผ ์ฐธ์กฐํ์ฌ ์๋์ ๊ฐ์ด ์์ฑํ์ธ์:
```py
>>> def wandb_hp_space(trial):
... return {
... "method": "random",
... "metric": {"name": "objective", "goal": "minimize"},
... "parameters": {
... "learning_rate": {"distribution": "uniform", "min": 1e-6, "max": 1e-4},
... "per_device_train_batch_size": {"values": [16, 32, 64, 128]},
... },
... }
```
`model_init` ํจ์๋ฅผ ์ ์ํ๊ณ ์ด๋ฅผ [`Trainer`]์ ์ ๋ฌํ์ธ์. ์๋๋ ๊ทธ ์์์
๋๋ค.
```py
>>> def model_init(trial):
... return AutoModelForSequenceClassification.from_pretrained(
... model_args.model_name_or_path,
... from_tf=bool(".ckpt" in model_args.model_name_or_path),
... config=config,
... cache_dir=model_args.cache_dir,
... revision=model_args.model_revision,
... token=True if model_args.use_auth_token else None,
... )
```
์๋์ ๊ฐ์ด `model_init` ํจ์, ํ๋ จ ์ธ์, ํ๋ จ ๋ฐ ํ
์คํธ ๋ฐ์ดํฐ์
, ๊ทธ๋ฆฌ๊ณ ํ๊ฐ ํจ์๋ฅผ ์ฌ์ฉํ์ฌ [`Trainer`]๋ฅผ ์์ฑํ์ธ์:
```py
>>> trainer = Trainer(
... model=None,
... args=training_args,
... train_dataset=small_train_dataset,
... eval_dataset=small_eval_dataset,
... compute_metrics=compute_metrics,
... tokenizer=tokenizer,
... model_init=model_init,
... data_collator=data_collator,
... )
```
ํ์ดํผํ๋ผ๋ฏธํฐ ํ์์ ํธ์ถํ๊ณ , ์ต์ ์ ์ํ ๋งค๊ฐ๋ณ์๋ฅผ ๊ฐ์ ธ์ค์ธ์. ๋ฐฑ์๋๋ `"optuna"`/`"sigopt"`/`"wandb"`/`"ray"` ์ค์์ ์ ํํ ์ ์์ต๋๋ค. ๋ฐฉํฅ์ `"minimize"` ๋๋ `"maximize"` ์ค ์ ํํ๋ฉฐ, ๋ชฉํ๋ฅผ ์ต์ํํ ๊ฒ์ธ์ง ์ต๋ํํ ๊ฒ์ธ์ง๋ฅผ ๊ฒฐ์ ํฉ๋๋ค.
์์ ๋ง์ compute_objective ํจ์๋ฅผ ์ ์ํ ์ ์์ต๋๋ค. ๋ง์ฝ ์ด ํจ์๋ฅผ ์ ์ํ์ง ์์ผ๋ฉด, ๊ธฐ๋ณธ compute_objective๊ฐ ํธ์ถ๋๊ณ , f1๊ณผ ๊ฐ์ ํ๊ฐ ์งํ์ ํฉ์ด ๋ชฉํฏ๊ฐ์ผ๋ก ๋ฐํ๋ฉ๋๋ค.
```py
>>> best_trial = trainer.hyperparameter_search(
... direction="maximize",
... backend="optuna",
... hp_space=optuna_hp_space,
... n_trials=20,
... compute_objective=compute_objective,
... )
```
## DDP ๋ฏธ์ธ ์กฐ์ ์ ์ํ ํ์ดํผํ๋ผ๋ฏธํฐ ํ์ [[hyperparameter-search-for-ddp-finetune]]
ํ์ฌ, DDP(Distributed Data Parallelism; ๋ถ์ฐ ๋ฐ์ดํฐ ๋ณ๋ ฌ์ฒ๋ฆฌ)๋ฅผ ์ํ ํ์ดํผํ๋ผ๋ฏธํฐ ํ์์ optuna์ sigopt์์ ๊ฐ๋ฅํฉ๋๋ค. ์ต์์ ํ๋ก์ธ์ค๊ฐ ํ์ดํผํ๋ผ๋ฏธํฐ ํ์ ๊ณผ์ ์ ์์ํ๊ณ ๊ทธ ๊ฒฐ๊ณผ๋ฅผ ๋ค๋ฅธ ํ๋ก์ธ์ค์ ์ ๋ฌํฉ๋๋ค.
| 0
|
hf_public_repos/transformers/docs/source
|
hf_public_repos/transformers/docs/source/ko/big_models.md
|
<!--Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
the License. You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
specific language governing permissions and limitations under the License.
โ ๏ธ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
rendered properly in your Markdown viewer.
-->
# ํฐ ๋ชจ๋ธ ์ธ์คํด์คํ [[instantiating-a-big-model]]
๋งค์ฐ ํฐ ์ฌ์ ํ๋ จ๋ ๋ชจ๋ธ์ ์ฌ์ฉํ๋ ค๋ฉด, RAM ์ฌ์ฉ์ ์ต์ํํด์ผ ํ๋ ๊ณผ์ ๊ฐ ์์ต๋๋ค. ์ผ๋ฐ์ ์ธ PyTorch ์ํฌํ๋ก์ฐ๋ ๋ค์๊ณผ ๊ฐ์ต๋๋ค:
1. ๋ฌด์์ ๊ฐ์ค์น๋ก ๋ชจ๋ธ์ ์์ฑํฉ๋๋ค.
2. ์ฌ์ ํ๋ จ๋ ๊ฐ์ค์น๋ฅผ ๋ถ๋ฌ์ต๋๋ค.
3. ์ฌ์ ํ๋ จ๋ ๊ฐ์ค์น๋ฅผ ๋ฌด์์ ๋ชจ๋ธ์ ์ ์ฉํฉ๋๋ค.
1๋จ๊ณ์ 2๋จ๊ณ ๋ชจ๋ ๋ชจ๋ธ์ ์ ์ฒด ๋ฒ์ ์ ๋ฉ๋ชจ๋ฆฌ์ ์ ์ฌํด์ผ ํ๋ฉฐ, ๋๋ถ๋ถ ๋ฌธ์ ๊ฐ ์์ง๋ง ๋ชจ๋ธ์ด ๊ธฐ๊ฐ๋ฐ์ดํธ๊ธ์ ์ฉ๋์ ์ฐจ์งํ๊ธฐ ์์ํ๋ฉด ๋ณต์ฌ๋ณธ 2๊ฐ๊ฐ RAM์ ์ด๊ณผํ์ฌ ๋ฉ๋ชจ๋ฆฌ ๋ถ์กฑ ์ด์๋ฅผ ์ผ๊ธฐํ ์ ์์ต๋๋ค. ๋ ์ฌ๊ฐํ ๋ฌธ์ ๋ ๋ถ์ฐ ํ์ต์ ์ํด `torch.distributed`๋ฅผ ์ฌ์ฉํ๋ ๊ฒฝ์ฐ, ํ๋ก์ธ์ค๋ง๋ค ์ฌ์ ํ๋ จ๋ ๋ชจ๋ธ์ ๋ก๋ํ๊ณ ๋ณต์ฌ๋ณธ์ 2๊ฐ์ฉ RAM์ ์ ์ฅํ๋ค๋ ๊ฒ์
๋๋ค.
<Tip>
๋ฌด์์๋ก ์์ฑ๋ ๋ชจ๋ธ์ "๋น์ด ์๋" (์ฆ ๊ทธ๋ ๋ฉ๋ชจ๋ฆฌ์ ์๋ ๊ฒ์ผ๋ก ์ด๋ค์ง) ํ
์๋ก ์ด๊ธฐํ๋๋ฉฐ ๋ฉ๋ชจ๋ฆฌ ๊ณต๊ฐ์ ์ฐจ์งํฉ๋๋ค. ์ด๊ธฐํ๋ ๋ชจ๋ธ/ํ๋ผ๋ฏธํฐ์ ์ข
๋ฅ์ ์ ํฉํ ๋ถํฌ(์: ์ ๊ท ๋ถํฌ)์ ๋ฐ๋ฅธ ๋ฌด์์ ์ด๊ธฐํ๋ ๊ฐ๋ฅํ ํ ๋น ๋ฅด๊ฒ ํ๊ธฐ ์ํด ์ด๊ธฐํ๋์ง ์์ ๊ฐ์ค์น์ ๋ํด 3๋จ๊ณ ์ดํ์๋ง ์ํ๋ฉ๋๋ค!
</Tip>
์ด ์๋ด์์์๋ Transformers๊ฐ ์ด ๋ฌธ์ ๋ฅผ ํด๊ฒฐํ๊ธฐ ์ํด ์ ๊ณตํ๋ ์๋ฃจ์
์ ์ดํด๋ด
๋๋ค. ์ฃผ์ํ ์ ์ ์์ง ํ๋ฐํ ๊ฐ๋ฐ ์ค์ธ ๋ถ์ผ์ด๋ฏ๋ก ์ฌ๊ธฐ์ ์ค๋ช
ํ๋ API๊ฐ ์์ผ๋ก ์ฝ๊ฐ ๋ณ๊ฒฝ๋ ์ ์๋ค๋ ๊ฒ์
๋๋ค.
## ์ค๋ฉ๋ ์ฒดํฌํฌ์ธํธ [[sharded-checkpoints]]
4.18.0 ๋ฒ์ ์ดํ, 10GB ์ด์์ ๊ณต๊ฐ์ ์ฐจ์งํ๋ ๋ชจ๋ธ ์ฒดํฌํฌ์ธํธ๋ ์๋์ผ๋ก ์์ ์กฐ๊ฐ๋ค๋ก ์ค๋ฉ๋ฉ๋๋ค. `model.save_pretrained(save_dir)`๋ฅผ ์คํํ ๋ ํ๋์ ๋จ์ผ ์ฒดํฌํฌ์ธํธ๋ฅผ ๊ฐ์ง๊ฒ ๋ ๋์ , ์ฌ๋ฌ ๋ถ๋ถ ์ฒดํฌํฌ์ธํธ(๊ฐ๊ฐ์ ํฌ๊ธฐ๋ 10GB ๋ฏธ๋ง)์ ๋งค๊ฐ๋ณ์ ์ด๋ฆ์ ํด๋น ํ์ผ์ ๋งคํํ๋ ์ธ๋ฑ์ค๊ฐ ์์ฑ๋ฉ๋๋ค.
`max_shard_size` ๋งค๊ฐ๋ณ์๋ก ์ค๋ฉ ์ ์ต๋ ํฌ๊ธฐ๋ฅผ ์ ์ดํ ์ ์์ผ๋ฏ๋ก, ์ด ์์ ๋ฅผ ์ํด ์ค๋ ํฌ๊ธฐ๊ฐ ์์ ์ผ๋ฐ ํฌ๊ธฐ์ ๋ชจ๋ธ์ ์ฌ์ฉํ๊ฒ ์ต๋๋ค: ์ ํต์ ์ธ BERT ๋ชจ๋ธ์ ์ฌ์ฉํด ๋ด
์๋ค.
```py
from transformers import AutoModel
model = AutoModel.from_pretrained("bert-base-cased")
```
[`~PreTrainedModel.save_pretrained`]์ ์ฌ์ฉํ์ฌ ๋ชจ๋ธ์ ์ ์ฅํ๋ฉด, ๋ชจ๋ธ์ ๊ตฌ์ฑ๊ณผ ๊ฐ์ค์น๊ฐ ๋ค์ด์๋ ๋ ๊ฐ์ ํ์ผ์ด ์๋ ์ ํด๋๊ฐ ์์ฑ๋ฉ๋๋ค:
```py
>>> import os
>>> import tempfile
>>> with tempfile.TemporaryDirectory() as tmp_dir:
... model.save_pretrained(tmp_dir)
... print(sorted(os.listdir(tmp_dir)))
['config.json', 'pytorch_model.bin']
```
์ด์ ์ต๋ ์ค๋ ํฌ๊ธฐ๋ฅผ 200MB๋ก ์ฌ์ฉํด ๋ด
์๋ค:
```py
>>> with tempfile.TemporaryDirectory() as tmp_dir:
... model.save_pretrained(tmp_dir, max_shard_size="200MB")
... print(sorted(os.listdir(tmp_dir)))
['config.json', 'pytorch_model-00001-of-00003.bin', 'pytorch_model-00002-of-00003.bin', 'pytorch_model-00003-of-00003.bin', 'pytorch_model.bin.index.json']
```
๋ชจ๋ธ์ ๊ตฌ์ฑ์ ๋ํด, ์ธ ๊ฐ์ ๋ค๋ฅธ ๊ฐ์ค์น ํ์ผ๊ณผ ํ๋ผ๋ฏธํฐ ์ด๋ฆ๊ณผ ํด๋น ํ์ผ์ ๋งคํ์ด ํฌํจ๋ `index.json` ํ์ผ์ ๋ณผ ์ ์์ต๋๋ค. ์ด๋ฌํ ์ฒดํฌํฌ์ธํธ๋ [`~PreTrainedModel.from_pretrained`] ๋ฉ์๋๋ฅผ ์ฌ์ฉํ์ฌ ์์ ํ ๋ค์ ๋ก๋ํ ์ ์์ต๋๋ค:
```py
>>> with tempfile.TemporaryDirectory() as tmp_dir:
... model.save_pretrained(tmp_dir, max_shard_size="200MB")
... new_model = AutoModel.from_pretrained(tmp_dir)
```
ํฐ ๋ชจ๋ธ์ ๊ฒฝ์ฐ ์ด๋ฌํ ๋ฐฉ์์ผ๋ก ์ฒ๋ฆฌํ๋ ์ฃผ๋ ์ฅ์ ์ ์์์ ๋ณด์ฌ์ค ํ๋ฆ์ 2๋จ๊ณ์์, ๊ฐ ์ค๋๊ฐ ์ด์ ์ค๋ ๋ค์์ ๋ก๋๋๋ฏ๋ก ๋ฉ๋ชจ๋ฆฌ ์ฌ์ฉ๋์ด ๋ชจ๋ธ ํฌ๊ธฐ์ ๊ฐ์ฅ ํฐ ์ค๋์ ํฌ๊ธฐ๋ฅผ ์ด๊ณผํ์ง ์๋๋ค๋ ์ ์
๋๋ค.
์ด ์ธ๋ฑ์ค ํ์ผ์ ํค๊ฐ ์ฒดํฌํฌ์ธํธ์ ์๋์ง, ๊ทธ๋ฆฌ๊ณ ํด๋น ๊ฐ์ค์น๊ฐ ์ด๋์ ์ ์ฅ๋์ด ์๋์ง๋ฅผ ๊ฒฐ์ ํ๋ ๋ฐ ์ฌ์ฉ๋ฉ๋๋ค. ์ด ์ธ๋ฑ์ค๋ฅผ json๊ณผ ๊ฐ์ด ๋ก๋ํ๊ณ ๋์
๋๋ฆฌ๋ฅผ ์ป์ ์ ์์ต๋๋ค:
```py
>>> import json
>>> with tempfile.TemporaryDirectory() as tmp_dir:
... model.save_pretrained(tmp_dir, max_shard_size="200MB")
... with open(os.path.join(tmp_dir, "pytorch_model.bin.index.json"), "r") as f:
... index = json.load(f)
>>> print(index.keys())
dict_keys(['metadata', 'weight_map'])
```
๋ฉํ๋ฐ์ดํฐ๋ ํ์ฌ ๋ชจ๋ธ์ ์ด ํฌ๊ธฐ๋ง ํฌํจ๋ฉ๋๋ค. ์์ผ๋ก ๋ค๋ฅธ ์ ๋ณด๋ฅผ ์ถ๊ฐํ ๊ณํ์
๋๋ค:
```py
>>> index["metadata"]
{'total_size': 433245184}
```
๊ฐ์ค์น ๋งต์ ์ด ์ธ๋ฑ์ค์ ์ฃผ์ ๋ถ๋ถ์ผ๋ก, ๊ฐ ๋งค๊ฐ๋ณ์ ์ด๋ฆ(PyTorch ๋ชจ๋ธ `state_dict`์์ ๋ณดํต ์ฐพ์ ์ ์๋)์ ํด๋น ํ์ผ์ ๋งคํํฉ๋๋ค:
```py
>>> index["weight_map"]
{'embeddings.LayerNorm.bias': 'pytorch_model-00001-of-00003.bin',
'embeddings.LayerNorm.weight': 'pytorch_model-00001-of-00003.bin',
...
```
๋ง์ฝ [`~PreTrainedModel.from_pretrained`]๋ฅผ ์ฌ์ฉํ์ง ์๊ณ ๋ชจ๋ธ ๋ด์์ ์ด๋ฌํ ์ค๋ฉ๋ ์ฒดํฌํฌ์ธํธ๋ฅผ ์ง์ ๊ฐ์ ธ์ค๋ ค๋ฉด (์ ์ฒด ์ฒดํฌํฌ์ธํธ๋ฅผ ์ํด `model.load_state_dict()`๋ฅผ ์ํํ๋ ๊ฒ์ฒ๋ผ), [`~modeling_utils.load_sharded_checkpoint`]๋ฅผ ์ฌ์ฉํด์ผ ํฉ๋๋ค.
```py
>>> from transformers.modeling_utils import load_sharded_checkpoint
>>> with tempfile.TemporaryDirectory() as tmp_dir:
... model.save_pretrained(tmp_dir, max_shard_size="200MB")
... load_sharded_checkpoint(model, tmp_dir)
```
## ์ (ไฝ)๋ฉ๋ชจ๋ฆฌ ๋ก๋ฉ [[low-memory-loading]]
์ค๋ฉ๋ ์ฒดํฌํฌ์ธํธ๋ ์์์ ์ธ๊ธํ ์์
ํ๋ฆ์ 2๋จ๊ณ์์ ๋ฉ๋ชจ๋ฆฌ ์ฌ์ฉ๋์ ์ค์ด์ง๋ง, ์ (ไฝ)๋ฉ๋ชจ๋ฆฌ ์ค์ ์์ ๋ชจ๋ธ์ ์ฌ์ฉํ๊ธฐ ์ํด ์ฐ๋ฆฌ์ Accelerate ๋ผ์ด๋ธ๋ฌ๋ฆฌ๋ฅผ ๊ธฐ๋ฐ์ผ๋ก ํ ๋๊ตฌ๋ฅผ ํ์ฉํ๋ ๊ฒ์ด ์ข์ต๋๋ค.
์์ธํ ์ฌํญ์ ๋ค์ ๊ฐ์ด๋๋ฅผ ์ฐธ์กฐํด์ฃผ์ธ์: [Accelerate๋ก ๋๊ท๋ชจ ๋ชจ๋ธ ๊ฐ์ ธ์ค๊ธฐ (์๋ฌธ)](../en/main_classes/model#large-model-loading)
| 0
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.