bidWdZddlZddlmZddlmZddlmZddlZddl Z ddl m Z m Z ddl mZddlmZd d lmZd d lmZmZmZmZmZmZmZmZmZmZmZmZm Z m!Z!m"Z"d d l#m$Z$d d l%m&Z&m'Z'm(Z(m)Z)eeeeeeeee e!eieee"edZ*e)j+e,Z-dZ.iddddddddddddddddd d!d"d#d$d%d&d'd(d)d*d+d,d-d.d/Z/gd0Z0d1Z1d2Z2d3Z3d4Z4d5Z5d6Z6eGd7d8Z7d9Z8dAd;Z9d<Z:d=Z;d>ZdS)Bz'Configuration base class and utilities.N) dataclass)Path)Any)is_offline_mode model_info)OfflineModeIsEnabled)HFValidationError) __version__),MODEL_FOR_AUDIO_CLASSIFICATION_MAPPING_NAMES!MODEL_FOR_CAUSAL_LM_MAPPING_NAMESMODEL_FOR_CTC_MAPPING_NAMES,MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES*MODEL_FOR_IMAGE_SEGMENTATION_MAPPING_NAMES*MODEL_FOR_IMAGE_TEXT_TO_TEXT_MAPPING_NAMES!MODEL_FOR_MASKED_LM_MAPPING_NAMES(MODEL_FOR_OBJECT_DETECTION_MAPPING_NAMES*MODEL_FOR_QUESTION_ANSWERING_MAPPING_NAMES,MODEL_FOR_SEQ_TO_SEQ_CAUSAL_LM_MAPPING_NAMES/MODEL_FOR_SEQUENCE_CLASSIFICATION_MAPPING_NAMES(MODEL_FOR_SPEECH_SEQ_2_SEQ_MAPPING_NAMES0MODEL_FOR_TABLE_QUESTION_ANSWERING_MAPPING_NAMES,MODEL_FOR_TOKEN_CLASSIFICATION_MAPPING_NAMES6MODEL_FOR_ZERO_SHOT_IMAGE_CLASSIFICATION_MAPPING_NAMES) ParallelMode)is_datasets_availableis_tokenizers_availableis_torch_availablelogging)text-generationimage-classificationimage-segmentation fill-maskobject-detectionquestion-answeringtext2text-generationtext-classificationtable-question-answeringtoken-classificationaudio-classificationautomatic-speech-recognitionzzero-shot-image-classificationzimage-text-to-textz r#zMasked Language Modelingr!zImage Classificationr"zImage Segmentationzmultiple-choicezMultiple Choicer$zObject Detectionr%zQuestion Answering summarization Summarizationr(zTable Question Answeringr'zText Classificationr zCausal Language Modelingr&z&Sequence-to-sequence Language Modelingr)zToken Classification translation Translationzzero-shot-classificationzZero Shot Classificationr+zAutomatic Speech Recognitionr*zAudio Classification) accuracybleuf1matthews_correlationpearsonr precisionrecallrouge sacrebleu spearmanrwerc>|gSt|tr|gS|SN) isinstancestr)objs Q/root/projects/butler/venv/lib/python3.11/site-packages/transformers/modelcard.py_listifyrAss* { C  u  c||St|tr|g}d|D}t|dkr|S|||<|S)Ncg|]}||Sr<.0vs r@ z*_insert_values_as_list..s 1 1 1A1=a===rBr)r=r>len)metadatanamevaluess r@_insert_values_as_listrN|s\ ~&# 1 1 1 1 1F 6{{aHTN OrBc |iSi}|D]z}|ddtvr,|||dd<]|dkr||d<{|S)N _rouge1r7)lowerreplace METRIC_TAGS) eval_resultsresultkeys r@#infer_metric_tags_from_eval_resultsrYs F"" 99;;  sC ( (K 7 747F399;;&&sC00 1 1 YY[[H $ $!F7O MrBc||S|||<|Sr<rE)rKrLvalues r@ _insert_valuer\s }HTN OrBcVtsdSddlm}m}t |||fS)NFr)DatasetIterableDataset)rdatasetsr^r_r=)datasetr^r_s r@ is_hf_datasetrbsB " "u11111111 g9 : ::rBcg}|D]F}t|ttfr|t|z }1||G|Sr<)rMr=tuplelistappend)mappingrWrHs r@_get_mapping_valuesrhsb F ^^   a% ' '  d1gg FF MM!     MrBc eZdZUeed<dZeeezdzed<dZedzed<dZeeezdzed<dZ edzed<dZ eeezdzed<dZ eeezdzed<dZ eeezdzed <dZ eeezdzed <dZeeefdzed <dZeeefdzed <dZeedzed <dZeeefdzed<dZedzed<dZdZdZdZe ddZdS)TrainingSummary model_nameNlanguagelicensetagsfinetuned_fromtasksra dataset_tags dataset_argsdataset_metadatarV eval_lineshyperparameterstrainersourcecJ|jts|jt|jdkrn t |j}|jD]&}|dr|dd|_'dS#tjttf$rYdSwxYwdSdSdSdS)Nrzlicense:) rmrrorJrrn startswithhttpx HTTPErrorr r)selfinfotags r@ __post_init__zTrainingSummary.__post_init__s L #%% !#/D'((1,, !$"5669//C~~j11/'*122w //O%68LM     ! //,,sAA::BBcd|ji}t|j}t|j}t|j}t|j}t |t |kr&|dgt |t |z zz}tt||tt||}tt||}dt|j D} g|d<t | dkrt dkr|gSt | dkrddi} t dkrddifd| D} | D]\} } i} | | | | d| d<| ?| | i}| | d|| d<|| || | dd <t |dkrIg| d <| D]/\}}| d  |||j |d 0d| vr$d| vr d | vr|d | td | |gS) NrLc<i|]}|tv |t|SrE)TASK_TAG_TO_NAME_MAPPING)rGtasks r@ z6TrainingSummary.create_model_index..s3   59UY]uUuUuD*40UuUuUurBresultsrc"g|] }D]}||f SrErE)rGtask_tagds_tagdataset_mappings r@rIz6TrainingSummary.create_model_index..s-kkkH[jkkQWh/kkkkrB)rLtyperraargsmetrics)rLrr[zLDropping the following result as it does not have all the necessary fields: )rkrArarqrrrsrJdictziprpgetitemsrfrVloggerr~)r}metric_mapping model_index dataset_namesrqrrrsdataset_arg_mappingdataset_metadata_mapping task_mappingall_possibilitiesrrrWrK metric_tag metric_namers @r@create_model_indexz"TrainingSummary.create_model_indexst/ !..  122  122 #D$9:: |  s<00 0 0'4&C 4E4ELHYHY4Y*ZZLs<??@@"3|\#B#BCC#'L:J(K(K#L#L   =Edj=Q=Q   "$ I |   ! !c/&:&:a&?&?= |   ! ! $""Q&&$&y!/=/C/C/E/E+J 9%,,$/$.%)%6{%CI$7$7IrJrnrqrekeysr)r}rrKs r@create_metadatazTrainingSummary.create_metadatas>JJ)(J NN 9dlCC   *z$:Ms/S/S *X[\`\oXpXpstXtXt$X|T=PQQH)(FDIFF)(J@QRR)(ItNDWDWDYDY?Z?Z[["&"9"9."I"IrBcd}tj|d}t|dkrd|d}|jdkr |t z }|d|jdz }|j|d z }n|d |jd |jd z }|j2t|jtrt|jdkr|d z }nt|jtr|d|jdz }nt|jttfr-t|jdkr|d|jddz }nA|d d|jddDd|jddzz }|j>|dz }|d d|jDz }|dz }|dz }|dz }|dz }|dz }|dz }|jD|dz }|d d|jDz }|dz }n|d z }|j!|d!z }|t%|jz }|dz }|d"z }|d#t&dz }|jdkr t)rddl}|d$|jdz }t-rddl}|d%|jdz }t1rddl}|d&|jdz }|S)'NF) sort_keysrz--- rvz # z z'This model was trained from scratch on z'This model is a fine-tuned version of [z](https://huggingface.co/z) on zan unknown dataset.the z dataset.r z, cg|]}d|S)rrE)rGdss r@rIz1TrainingSummary.to_model_card..<sGGGr{b{{GGGrBz and the z datasets.z: It achieves the following results on the evaluation set:  c>g|]\}}d|dt|Sz- z:  _maybe_roundrGrLr[s r@rIz1TrainingSummary.to_model_card..As6$u$u$uKDRW%G$%G%G,u2E2E%G%G$u$u$urBz/ ## Model description More information needed z9 ## Intended uses & limitations More information needed z: ## Training and evaluation data More information needed z ## Training procedure z ### Training hyperparameters z: The following hyperparameters were used during training: c$g|] \}}d|d|SrrErs r@rIz1TrainingSummary.to_model_card..Ls-$j$j$jkdE%9$%9%9%%9%9$j$j$jrBz More information needed z ### Training results z ### Framework versions z- Transformers z - Pytorch z - Datasets z - Tokenizers )yamldumprrJrwAUTOGENERATED_TRAINER_COMMENTrkrorar=rer>rdjoinrVrrurtmake_markdown_tabler rtorchrr`r tokenizers)r} model_cardrKrr`rs r@ to_model_cardzTrainingSummary.to_model_cards 9T1133uEEE x==1  0000J ;) # # 7 7J2T_2222   & C CJJ ^(^^CGCV^^^ J < Jt|T$B$B s4> >J $ J44 7 7777 ;) # #(:(<(< # LLL rhasattrmodelrospathisdirr __class____name__ TASK_MAPPINGrrhrr output_dirrLrJr=rfparse_log_historystate log_history$extract_hyperparameters_from_trainer)clsrvrlrmrnrkrorprqrsrarr one_dataset default_tagmodel_class_namerrgrQrtrVrus r@ from_trainerzTrainingSummary.from_trainerhs/ /6.B.Ng**T[Ti  % % =<+?3JUXYdYjUkUk(l(l'm$'$/=L'$/$;#> #GMM'-"6"DEE #%]1?N =&}6? !-!3!3!5!5 ! ! g#':7'C'CCC E  gl566;J z??a  'J <,-DD c " " 2t/G'G'G23DD %T 1 1 KK0 1 1 1&7 8Q&R&R#:|>wGGs!)%%-%!+    rB) NNNNNNNNNN)r __module__ __qualname__r>__annotations__rlrermrnrorprarqrrrsrrrVfloatrtrurwrrrr classmethodrrErBr@rjrjsOOO'+HcDIo$+++GS4Z#'D#S /D '''!%NC$J%%%$(E3c?T !(((&*GS49_t #***+/L#S /D(///+/L#S /D(///.2d38nt+222,0L$sEz"T)000#'JS D '''-1OT#s(^d*111"FC$J""" ===~ HHHT K K K [K K K rBrjc\d}|t|kr,d||vr"|dz }|t|kr d||v"|t|kr=|dz}|dkrd||vr|dz}|dkr d||v|dkr dd||fSdS||}g}d}t|D]F}d||vr||d}d||vr ||}|d d}|d d}|d d} |d d}|d d}|dd}||| d} |D]R\} } | dkr| | d<| d} dd| ddD}| | |<S|| Ht|dz }|dkrd||vr|dz}|dkr d||v|dkrwi}||D]U\}}|d}|dvr7dd|dD}|||<V|||fS||dfS)zd Parse the `log_history` of a Trainer to get the intermediate and final evaluation results. r train_runtimer eval_lossN)NNNzNo logloss total_flosepochstep eval_runtimeeval_samples_per_secondeval_steps_per_second)z Training LossEpochStepzValidation LossrQrPc6g|]}|SrE capitalizerGparts r@rIz%parse_log_history..s"$N$N$N4T__%6%6$N$N$NrBeval_)runtimesamples_per_secondsteps_per_secondrrc6g|]}|SrErrs r@rIz%parse_log_history..s"+Y+Y+Y$DOO,=,=+Y+Y+YrB) rJrangecopypoprrrrf removeprefix)ridx train_loglines training_lossirrQrrrMkrHsplitsrLrVrXr[camel_cased_keys r@rrs< C K  _K>> 1HCQhh;k#.>>> !88{3// /##C I EM 3ZZ!! [^ # #'N62M +a. ( (!!n))++G L$//AKK..E;;vt,,D ND11A 5t<r)rHdecimalss r@rrs!U#CFFLL$5$5 6 6 : :s3q66<.s:___tq!aC1s1vv:>$::___rBr| )rr)rM col_widthsvalues_with_spaces r@_regular_table_liner s9__s6S]G^G^___ 77$ % % --rBcJd|D}d|dzS)Nc$g|] }dd|zzdzS)z|:-:rE)rGrs r@rIz&_second_table_line..s% 7 7 7qdS1Wns" 7 7 7rBrr )r)r rMs r@_second_table_liners* 7 7J 7 7 7F 776??U ""rBc |t|dkrdSd|dD}|D]a}|D]J\}}||tt|krtt|||<Kbtt |dt |}|tt |z }|D]P}|td|Dt |z }Q|S)zC Create a nice Markdown table from the results in `lines`. NrrcHi|]}|tt| SrE)rJr>)rGrXs r@rz'make_markdown_table.. s&999#s3s88}}999rBc,g|]}t|SrErrFs r@rIz'make_markdown_table..s%M%M%M!l1oo%M%M%MrB)rJrrr rerrMr)rr linerXr[tables r@rrsT }E ar99a999J;;**,, ; ;JC#\%%8%8!9!999"%l5&9&9":": 3 ; U1X]]__ 5 5tJ.s'PPPqq'',22PPPrBz multi-GPUdistributed_typer num_devicesgradient_accumulation_stepsrtotal_train_batch_sizertotal_eval_batch_sizez!No additional optimizer argumentsadamzUse z with betas=(,z) and epsilon=z and optimizer_args= optimizerz and the args are: lr_scheduler_typeglr_scheduler_warmup_stepsrtraining_steps num_epochsz Native AMPmixed_precision_traininglabel_smoothing_factor)_TRAINING_ARGS_KEYSr parallel_moder NOT_PARALLELNOT_DISTRIBUTED DISTRIBUTEDr[ world_sizer rroptim optim_argsrS adam_beta1 adam_beta2 adam_epsilonr& warmup_steps max_stepsnum_train_epochsfp16r+)rvrur!r"optimizer_nameoptimizer_argss` r@rrsPPPP|7P34|##,3L,B())(/ (E %|C6B23|*c114;L4W01 rB)r)?__doc__r dataclassesrpathlibrtypingrr{rhuggingface_hubrrhuggingface_hub.errorsrhuggingface_hub.utilsr rr models.auto.modeling_autor r rrrrrrrrrrrrr training_argsrutilsrrrrr get_loggerrrrrrUrArNrYr\rbrhrjrrr rrr,rrErBr@rHs.- !!!!!! 77777777777777333333"('''''9HD2@DHJ PHH$o'B$oFn$o&\D "  H % %! +2.(  *  . _ :01D2= :#$B 2!(          ;;; E E E E E E E  E P9&9&9&x ... ### &/////rB