h~SrSSKrSSKrSSKrSSKrSSKrSSKrSSKJrJ r SSK r SSK J r SSK JrJrJrJr SSKJr SSKJr SS KJr SS KJrJr SS KJrJr SS KJrJ r J!r!J"r"J#r#J$r$J%r%J&r&J'r'J(r(J)r)J*r*J+r+ SS K,J-r- \R\"\/5r0\ Rbr2SSK3r3\3Rhr2"SS5r6"SS5r7"SS5r8g!\5a N'f=f)zlProvides an API for writing protocol buffers to event files to be consumed by TensorBoard for visualization.N)OptionalUnion) CometLogger) append_pbtxtmake_mat make_spritemake_tsv)EventFileWriter)load_onnx_graph)load_openvino_graph) event_pb2 summary_pb2)Event SessionLog) audiocustom_scalars histogram histogram_rawhparamsimage image_boxesmeshpr_curve pr_curve_rawscalartextvideo)figure_to_imagec\rSrSrSrSrSrSSjrSSjrSSjr SS jr S r S r S r \RSSS .Sj5rSrg)DummyFileWriter1z8A fake file writer that writes nothing to the disk. cXlgN_logdir)selflogdirs M/var/www/fran/franai/venv/lib/python3.13/site-packages/tensorboardX/writer.py__init__DummyFileWriter.__init__5s cUR$z7Returns the directory where event file will be written.r%r's r) get_logdirDummyFileWriter.get_logdir8s ||r,Ncgr$r'eventstepwalltimes r) add_eventDummyFileWriter.add_event<r,cgr$r3)r'summary global_stepr7s r) add_summaryDummyFileWriter.add_summary?r:r,cgr$r3)r' graph_profiler7s r) add_graphDummyFileWriter.add_graphBr:r,cgr$r3)r'graphr7s r)add_onnx_graphDummyFileWriter.add_onnx_graphEr:r,cgr$r3r/s r)flushDummyFileWriter.flushHr:r,cgr$r3r/s r)closeDummyFileWriter.closeKr:r,cgr$r3r/s r)reopenDummyFileWriter.reopenNr:r,r=r7c## Uv g7fr$r3r'r=r7s r) use_metadataDummyFileWriter.use_metadataQs  sr%NNr$)__name__ __module__ __qualname____firstlineno____doc__r*r0r8r>rBrFrIrLrO contextlibcontextmanagerrT__static_attributes__r3r,r)r!r!1sS*.r,r!c\rSrSrSrSSjrSrSSjrSSjrSSjr SS jr SS jr S r S r S r\R SSS.Sj5rSrg) FileWriterVaWrites protocol buffers to event files to be consumed by TensorBoard. The `FileWriter` class provides a mechanism to create an event file in a given directory and add summaries and events to it. The class updates the file contents asynchronously. This allows a training program to call methods to add data to the file directly from the training loop, without slowing down training. c^[U5n[XX45TlU4Sjn[R"U5 0Tlg)aCreates a `FileWriter` and an event file. On construction the writer creates a new event file in `logdir`. The other arguments to the constructor control the asynchronous writes to the event file. Args: logdir: A string. Directory where event file will be written. max_queue: Integer. Size of the queue for pending events and summaries before one of the 'add' calls forces a flush to disk. Default is ten items. flush_secs: Number. How often, in seconds, to flush the pending events and summaries to disk. Default is every two minutes. filename_suffix: A string. Suffix added to all event filenames in the logdir directory. More details on filename construction in tensorboard.summary.writer.event_file_writer.EventFileWriter. c:>TRR5 gr$ event_writerrLr/sr)cleanup$FileWriter.__init__..cleanupys    # # %r,N)strr reatexitregister_default_metadata)r'r( max_queue flush_secsfilename_suffixrfs` r)r*FileWriter.__init__`s;*V+ z< &  !#r,c6URR5$r.)rer0r/s r)r0FileWriter.get_logdirs  ++--r,NcUc/URRS[R"55OUnUbX1lUcURRS5OUnUb[ U5UlUR RU5 g)aAdds an event to the event file. Args: event: An `Event` protocol buffer. step: Number. Optional global step value for training process to record with the event. walltime: float. Optional walltime to override the default (current) walltime (from time.time()) Nr7r=)rkgettime wall_timeintr6rer8r4s r)r8FileWriter.add_events  " " & &z499; ?   &O<@Lt%%))-8d  TEJ ##E*r,cP[R"US9nURXBU5 g)aAdds a `Summary` protocol buffer to the event file. This method wraps the provided summary in an `Event` protocol buffer and adds it to the event file. Args: summary: A `Summary` protocol buffer. global_step: Number. Optional global step value for training process to record with the summary. walltime: float. Optional walltime to override the default (current) walltime (from time.time()) )r<N)rrr8)r'r<r=r7r5s r)r>FileWriter.add_summarys 0 u84r,cUSnUSn[R"UR5S9nURUSU5 [R"SUR5S9n[R"US9nURUSU5 g)zAdds a `Graph` and step stats protocol buffer to the event file. Args: graph_profile: A `Graph` and step stats protocol buffer. walltime: float. Optional walltime to override the default (current) walltime (from time.time()) seconds after epoch rr graph_defNprofiler)tag run_metadata)tagged_run_metadata)rrSerializeToStringr8TaggedRunMetadata)r'rAr7rE stepstatsr5trms r)rBFileWriter.add_graphs|a !!$ %*A*A*CD udH-)))D)D)FHC8 udH-r,cn[R"UR5S9nURUSU5 gzAdds a `Graph` protocol buffer to the event file. Args: graph: A `Graph` protocol buffer. walltime: float. Optional walltime to override the default (current) _get_file_writerfrom time.time()) r{Nrrrr8r'rEr7r5s r)rFFileWriter.add_onnx_graph+%*A*A*CD udH-r,cn[R"UR5S9nURUSU5 grrrs r)add_openvino_graphFileWriter.add_openvino_graphrr,c8URR5 g)zqFlushes the event file to disk. Call this method to make sure that all pending events have been written to disk. N)rerIr/s r)rIFileWriter.flushs !r,c8URR5 g)zuFlushes the event file to disk and close the file. Call this method when you do not need the summary writer anymore. Nrdr/s r)rLFileWriter.closes !r,c8URR5 g)zReopens the EventFileWriter. Can be called after `close()` to add more events in the same directory. The events will go into a new events file. Does nothing if the EventFileWriter was not closed. N)rerOr/s r)rOFileWriter.reopens   "r,rQc#~# UR(aS5eUc UcS5eXS.UlUv 0Ulg7f)a(Context manager to temporarily set default metadata for all enclosed :meth:`add_event` calls. Args: global_step: Global step value to record walltime: Walltime to record (defaults to time.time()) Examples:: with writer.use_metadata(global_step=10): writer.add_event(event) z Default metadata is already set.Nz=At least one of `global_step` or `walltime` must be provided.rQ)rkrSs r)rTFileWriter.use_metadatasQ))M+MM)  #x'; K J K ;1rBrFrrIrLrOr\r]rTr^r3r,r)r`r`VsX$>.+. 5.$ . ."" #*.$$r,r`cp\rSrSrSrSSSSSSSSSS04 S \\S \\S \\S \\S \\S\\S\\S\\S\\ 4Sjjr Sr Sr Sr S]S\ \\\\\\44S\ \\4S\\S\\4SjjrS^S\S\\\4S\\S\\S\\S\\4 S jjrS]S!\S"\ \\4S\\S\\4S#jjrS$rS_S\S%\S\\S&\\S\\4 S'jjrS]S\S\\S\\4S(jjrS`S\S)\S\\S\\S*\\4 S+jjrSaS\S)\S\\S\\S*\\4 S,jjrSbS\S)\S-\S\\S\\S*\\S.\\\4S/jjrScS\S\\S0\\S\\4S1jjrSdS\S2\S\\S3\\\\4S\\S*\\4 S4jjrSeS\S5\S\\S6\\S\\4 S7jjrS]S\S8\S\\S\\4S9jjrS:rS;r SfS<jr!\"S=5r#SgS>\S?\S\\4S@jjr$ShS\S.\SA\S\\SB\\S\\4 SCjjr%ShS\SD\SE\SF\SG\SH\SI\S\\SB\\S\\4SJjjr&SiSK\\SL\SM\4SNjjr'SiSK\\SL\SM\4SOjjr(SP\ \\ \\444SQjr)SjS\SR\SS\ST\S\\S\\4 SUjjr*SVr+SWr,SXr-SYr.\/R`SSSZ.S[j5r1S\r2g)k SummaryWriteraWrites entries directly to event files in the logdir to be consumed by TensorBoard. The `SummaryWriter` class provides a high-level API to create an event file in a given directory and add summaries and events to it. The class updates the file contents asynchronously. This allows a training program to call methods to add data to the file directly from the training loop, without slowing down training. NrrrTdisabledr(comment purge_steprlrmrn write_to_disklog_dir comet_configc  HUbUcUnU(d`SSKn SSKJn U R5RS5n [R R SU S-U R5-U-5nXlX0l X@l XPl X`l Xpl XlSUlXlS=UlUlUR'5 Sn/n/nUS:a0UR)U5 UR)U*5 US -nUS:aM0USSS 2S/-U-Ul0Ul0Ulg) a%Creates a `SummaryWriter` that will write out events and summaries to the event file. Args: logdir: Save directory location. Default is runs/**CURRENT_DATETIME_HOSTNAME**, which changes after each run. Use hierarchical folder structure to compare between runs easily. e.g. pass in 'runs/exp1', 'runs/exp2', etc. for each new experiment to compare across them. comment: Comment logdir suffix appended to the default ``logdir``. If ``logdir`` is assigned, this argument has no effect. purge_step: When logging crashes at step :math:`T+X` and restarts at step :math:`T`, any events whose global_step larger or equal to :math:`T` will be purged and hidden from TensorBoard. Note that crashed and resumed experiments should have the same ``logdir``. max_queue: Size of the queue for pending events and summaries before one of the 'add' calls forces a flush to disk. Default is ten items. flush_secs: How often, in seconds, to flush the pending events and summaries to disk. Default is every two minutes. filename_suffix: Suffix added to all event filenames in the logdir directory. More details on filename construction in tensorboard.summary.writer.event_file_writer.EventFileWriter. write_to_disk: If pass `False`, SummaryWriter will not write to disk. comet_config: A comet config dictionary. Contains parameters that need to be passed to comet like workspace, project_name, api_key, disabled etc Examples:: from tensorboardX import SummaryWriter # create a summary writer with automatically generated folder name. writer = SummaryWriter() # folder location: runs/May04_22-14-54_s-MacBook-Pro.local/ # create a summary writer using the specified folder name. writer = SummaryWriter("my_experiment") # folder location: my_experiment # create a summary writer with comment appended. writer = SummaryWriter(comment="LR_0.1_BATCH_16") # folder location: runs/May04_22-14-54_s-MacBook-Pro.localLR_0.1_BATCH_16/ Nr)datetimez %b%d_%H-%M-%Sruns_g-q=g@xDg?)socketrnowstrftimeospathjoin gethostnamer(r _max_queue _flush_secs_filename_suffix_write_to_disk _comet_config _comet_loggerkwargs file_writer all_writers_get_file_writerappend default_bins scalar_dictrk)r'r(rrrlrmrnrrrrrr current_timevbuckets neg_bucketss r)r*SummaryWriter.__init__s*v  6>F  )#<<>22?CLWW\\ s*V-?-?-AAGKMF $#% /+)! /324+   $h NN1    r " HA$h("-3g=!#r,c SSKJn XR;a/URU'URURXC[ U"U5R 55/5 g)zyThis adds an entry to the self.scalar_dict datastructure with format {writer_id : [[timestamp, step, value], ...], ...}. rmake_npN)x2numrrrfloatsqueeze)r'r~ scalar_valuer= timestamprs r)__append_to_scalar_dict%SummaryWriter.__append_to_scalar_dictgsX # && &$&D  S ! $$ U7<+@+H+H+J%K L Nr,c UR(dO[URS9UlURR 5UR0UlUR$UR b URc[ SURURURURS.URD6UlURboURR[URSS95 URR[UR[[RS9S95 URR 5UR0UlUR$)z@Returns the default FileWriter instance. Recreates it if closed.r()r(rlrmrnz brain.Event:2)r6 file_version)status)r6 session_logr3)rr!r(rr0rr`rrrrrr8rrSTARTr/s r)rSummaryWriter._get_file_writerrs+"".dkkBD  $ 0 0 ; ; =t?O?OPD ## #    #t'7'7'?) 948OO595E5E:>:O:O 9-1KK  9D  *  **t_MO  **tJjN^N^<_`b $ 0 0 ; ; =t?O?OPD r,chURc[UR5UlUR$)z8Returns a comet logger instance. Recreates it if closed.)rrrr/s r)_get_comet_loggerSummaryWriter._get_comet_loggers-    %!,T-?-?!@D !!!r, hparam_dict metric_dictnamer=c[U5[Ld[U5[La [S5e[X5upVnU(d[ [ R "55n[ [RRURR5U5S9nURRU5 URRU5 URRU5 UR5HupURXU5 M SSS5 UR5R!XS9 g!,(df  N,=f)aAdd a set of hyperparameters to be compared in tensorboard. Args: hparam_dict: Each key-value pair in the dictionary is the name of the hyper parameter and it's corresponding value. metric_dict: Each key-value pair in the dictionary is the name of the metric and it's corresponding value. Note that the key used here should be unique in the tensorboard record. Otherwise the value you added by `add_scalar` will be displayed in hparam plugin. In most cases, this is unwanted. name: Personnalised name of the hparam session global_step: Current time step Examples:: from tensorboardX import SummaryWriter with SummaryWriter() as w: for i in range(5): w.add_hparams({'lr': 0.1*i, 'bsize': i}, {'hparam/accuracy': 10*i, 'hparam/loss': 10*i}) Expected result: .. image:: _static/img/tensorboard/add_hparam.png :scale: 50 % z1hparam_dict and metric_dict should be dictionary.rNr6)typedict TypeErrorrrhrtrrrrrr0r>items add_scalarrlog_parameters) r'rrrr=expssiseiw_hpkrs r) add_hparamsSummaryWriter.add_hparamssB  D (D,=T,IOP P 9 #tyy{#D "'',,t/?/?/J/J/Ld"S TX\    ( ( -    ( ( -    ( ( -#))+k2, U  // /N U Ts "A=E Er~rr7 display_namesummary_descriptioncUR5R[XXV5X45 UR5R XX#5 g)aAdd scalar data to summary. Args: tag: Data identifier scalar_value: Value to be logged in tensorboard global_step: Global step value to record walltime: Optional override default walltime (time.time()) of event display_name: The title of the plot. If empty string is passed, `tag` will be used. summary_description: The comprehensive text that will showed by clicking the information icon on TensorBoard. Examples:: from tensorboardX import SummaryWriter writer = SummaryWriter() x = range(100) for i in x: writer.add_scalar('y=2x', i * 2, i) writer.close() Expected result: .. image:: _static/img/tensorboard/add_scalar.png :scale: 50 % N)rr>rr log_metric)r'r~rr=r7rrs r)rSummaryWriter.add_scalarsBD ++ 3l H+ a  ++C|Yr,main_tagtag_scalar_dictcUc[R"5OUnUR5R5nUR5Hupg[R R [U5X5nXR;aURUn O[US9n XRU'U R[X5X45 URXX45 M UR5RX!US9 g)aAdds many scalar data to summary. Note that this function also keeps logged scalars in memory. In extreme case it explodes your RAM. Args: main_tag: The parent name for the tags tag_scalar_dict: Key-value pair storing the tag and corresponding values global_step: Global step value to record walltime: Optional override default walltime (time.time()) of event Examples:: from tensorboardX import SummaryWriter writer = SummaryWriter() r = 5 for i in range(100): writer.add_scalars('run_14h', {'xsinx':i*np.sin(i/r), 'xcosx':i*np.cos(i/r), 'tanx': np.tan(i/r)}, i) writer.close() # This call adds three values to the same scalar plot with the tag # 'run_14h' in TensorBoard's scalar section. Expected result: .. image:: _static/img/tensorboard/add_scalars.png :scale: 50 % Nrr)rtrr0rrrrrhrr`r>r%_SummaryWriter__append_to_scalar_dictr log_metrics) r'rrr=r7 fw_logdirr~rfw_tagfws r) add_scalarsSummaryWriter.add_scalarssF#+"2499;))+668 !0!6!6!8 CWW\\#i.(@F)))%%f-v.+-  ( NN6(9& 2  ( (k ="9  ,,_[,Yr,c[US5n[R"URU5 SSS5 0Ulg!,(df  N=f)zExports to the given path an ASCII file containing all the scalars written so far by this instance, with the following format: {writer_id : [[timestamp, step, value], ...], ...} The scalars saved by ``add_scalars()`` will be flushed after export. wN)openjsondumpr)r'rfs r)export_scalars_to_json$SummaryWriter.export_scalars_to_jsons8$_ IId&& *_s "? A valuesbinsc [U[5(aUS:Xa URnUR5R [ XXFS9X55 UR 5RX!U5 g)aAdd histogram to summary. Args: tag: Data identifier values: Values to build histogram global_step: Global step value to record bins: One of {'tensorflow','auto', 'fd', ...}. This determines how the bins are made. You can find other options in the `numpy reference `_. walltime: Optional override default walltime (time.time()) of event Examples:: from tensorboardX import SummaryWriter import numpy as np writer = SummaryWriter() for i in range(10): x = np.random.random(1000) writer.add_histogram('distribution centers', x + i, i) writer.close() Expected result: .. image:: _static/img/tensorboard/add_histogram.png :scale: 50 % tensorflow)max_binsN) isinstancerhrrr>rr log_histogram)r'r~rr=rr7rs r) add_histogramSummaryWriter.add_histogram saF dC T\%9$$D ++ c4 ;[ T  ..vKHr,c [U5[U5:wa [S5e[UUUUUUUU5n UR5R U U U 5 UR 5R XU S9 g)agAdds histogram with raw data. Args: tag: Data identifier min (float or int): Min value max (float or int): Max value num (int): Number of values sum (float or int): Sum of all values sum_squares (float or int): Sum of squares for all values bucket_limits (torch.Tensor, numpy.array): Upper value per bucket, note that the bucket_limits returned from `np.histogram` has one more element. See the comment in the following example. bucket_counts (torch.Tensor, numpy.array): Number of values per bucket global_step: Global step value to record walltime: Optional override default walltime (time.time()) of event Examples:: import numpy as np dummy_data = [] for idx, value in enumerate(range(30)): dummy_data += [idx + 0.001] * value values = np.array(dummy_data).astype(float).reshape(-1) counts, limits = np.histogram(values) sum_sq = values.dot(values) with SummaryWriter() as summary_writer: summary_writer.add_histogram_raw( tag='hist_dummy_data', min=values.min(), max=values.max(), num=len(values), sum=values.sum(), sum_squares=sum_sq, bucket_limits=limits[1:].tolist(), # <- note here. bucket_counts=counts.tolist(), global_step=0) z;len(bucket_limits) != len(bucket_counts), see the document.rN)len ValueErrorrrr>rlog_histogram_raw) r'r~minmaxnumsum sum_squares bucket_limits bucket_countsr=r7r<s r)add_histogram_rawSummaryWriter.add_histogram_rawIsd } ]!3 3Z[ [ # # # # + - -/ ++      223k2Rr, img_tensor dataformatsc[XUS9nURSRRnUR5R XcU5 UR 5R XqUS9 g)aAdd image data to summary. Note that this requires the ``pillow`` package. Args: tag: Data identifier img_tensor: An `uint8` or `float` Tensor of shape ` [channel, height, width]` where `channel` is 1, 3, or 4. The elements in img_tensor can either have values in [0, 1] (float32) or [0, 255] (uint8). Users are responsible to scale the data in the correct range/type. global_step: Global step value to record walltime: Optional override default walltime (time.time()) of event. dataformats: This parameter specifies the meaning of each dimension of the input tensor. Shape: img_tensor: Default is :math:`(3, H, W)`. You can use ``torchvision.utils.make_grid()`` to convert a batch of tensor into 3xHxW format or use ``add_images()`` and let us do the job. Tensor with :math:`(1, H, W)`, :math:`(H, W)`, :math:`(H, W, 3)` is also suitible as long as corresponding ``dataformats`` argument is passed. e.g. CHW, HWC, HW. Examples:: from tensorboardX import SummaryWriter import numpy as np img = np.zeros((3, 100, 100)) img[0] = np.arange(0, 10000).reshape(100, 100) / 10000 img[1] = 1 - np.arange(0, 10000).reshape(100, 100) / 10000 img_HWC = np.zeros((100, 100, 3)) img_HWC[:, :, 0] = np.arange(0, 10000).reshape(100, 100) / 10000 img_HWC[:, :, 1] = 1 - np.arange(0, 10000).reshape(100, 100) / 10000 writer = SummaryWriter() writer.add_image('my_image', img, 0) # If you have non-default dimension setting, set the dataformats argument. writer.add_image('my_image_HWC', img_HWC, 0, dataformats='HWC') writer.close() Expected result: .. image:: _static/img/tensorboard/add_image.png :scale: 50 % rrrN)rvalueencoded_image_stringrr>rlog_image_encoded)r'r~rr=r7rr<rs r) add_imageSummaryWriter.add_imagesfh[A&}}Q/55JJ ++ ( ,  223GS^2_r,c[U[5(a\UR5S:wa+UR5S:wa[S5 [S5 gSSKnUR "US5nSU-n[XUS9nURSRRn UR5RXU5 UR5RXUS 9 g![ a SSKnUR US5nNf=f) aAdd batched (4D) image data to summary. Besides passing 4D (NCHW) tensor, you can also pass a list of tensors of the same size. In this case, the ``dataformats`` should be `CHW` or `HWC`. Note that this requires the ``pillow`` package. Args: tag: Data identifier img_tensor: Image data The elements in img_tensor can either have values in [0, 1] (float32) or [0, 255] (uint8). Users are responsible to scale the data in the correct range/type. global_step: Global step value to record walltime: Optional override default walltime (time.time()) of event Shape: img_tensor: Default is :math:`(N, 3, H, W)`. If ``dataformats`` is specified, other shape will be accepted. e.g. NCHW or NHWC. Examples:: from tensorboardX import SummaryWriter import numpy as np img_batch = np.zeros((16, 3, 100, 100)) for i in range(16): img_batch[i, 0] = np.arange(0, 10000).reshape(100, 100) / 10000 / 16 * i img_batch[i, 1] = (1 - np.arange(0, 10000).reshape(100, 100) / 10000) / 16 * i writer = SummaryWriter() writer.add_images('my_image_batch', img_batch, 0) writer.close() Expected result: .. image:: _static/img/tensorboard/add_images.png :scale: 30 % CHWHWCzEA list of image is passed, but the dataformat is neither CHW nor HWC.zNothing is written.NrNrr)rlistupperprinttorchstackrnumpyrrrrr>rr) r'r~rr=r7rr$npr<rs r) add_imagesSummaryWriter.add_imagessV j$ ' '  "e+ 0A0A0Cu0L]^+,  5"[[Q7  +K[A&}}Q/55JJ ++ ( ,  223GS^2_ 5"XXj!4  5sC DD box_tensorlabelsc UbK[U[5(aU/n[U5URS:wa[R S5 Sn[ XU4XgS.UD6n U RSRRn UR5RXU5 UR5RXUS9 g)aAdd image and draw bounding boxes on the image. Args: tag: Data identifier img_tensor: Image data box_tensor: Box data (for detected objects) box should be represented as [x1, y1, x2, y2]. global_step: Global step value to record walltime: override default walltime (time.time()) of event labels: The strings to be show on each bounding box. Shape: img_tensor: Default is :math:`(3, H, W)`. It can be specified with ``dataformats`` argument. e.g. CHW or HWC box_tensor: (torch.Tensor, numpy.array, or string/blobname): NX4, where N is the number of boxes and each 4 elements in a row represents (xmin, ymin, xmax, ymax). Nrz@Number of labels do not equal to number of box, skip the labels.)rr+r)rrhrshapeloggerwarningrrrrrr>rr) r'r~rr*r=r7rr+rr<rs r)add_image_with_boxes"SummaryWriter.add_image_with_boxess6  &#&& 6{j..q11ab Z[5@[SY[&}}Q/55JJ ++ ( ,  223GS^2_r,rLc[U[5(aURU[X$5X5SS9 gURU[X$5X5SS9 g)aRender matplotlib figure into an image and add it to summary. Note that this requires the ``matplotlib`` package. Args: tag: Data identifier figure (matplotlib.pyplot.figure) or list of figures: Figure or a list of figures global_step: Global step value to record close: Flag to automatically close the figure walltime: Override default walltime (time.time()) of event NCHWrrN)rr!rr)r'r~figurer=rLr7s r) add_figureSummaryWriter.add_figure,sE$ fd # # NN3 > ciN j NN3 > chN ir, vid_tensorfpsc[XXFS9nURSRRnUR 5R XsU5 UR 5RXUS9 g)aAdd video data to summary. Note that this requires the ``moviepy`` package. Args: tag: Data identifier vid_tensor: Video data global_step: Global step value to record fps: Frames per second walltime: Optional override default walltime (time.time()) of event dataformats: Specify different permutation of the video tensor Shape: vid_tensor: :math:`(N, T, C, H, W)`. The values should lie in [0, 255] for type `uint8` or [0, 1] for type `float`. rrrN)rrrrrr>rr) r'r~r7r=r8r7rr<rs r) add_videoSummaryWriter.add_videoCse.F&}}Q/55JJ ++ ( ,  223GS^2_r, snd_tensor sample_ratecUR5R[XUS9X55 UR5R X$XS9 g)aAdd audio data to summary. Args: tag: Data identifier snd_tensor: Sound data global_step: Global step value to record sample_rate: sample rate in Hz walltime: Optional override default walltime (time.time()) of event Shape: snd_tensor: :math:`(L, C)`. The values should lie between [-1, 1]. Where `L` is the number of audio frames and `C` is the channel. Set channel equals to 2 for stereo. )r=rN)rr>rr log_audio)r'r~r<r=r=r7s r) add_audioSummaryWriter.add_audio`sD( ++ #{ ;[ T  **:C*Zr, text_stringcUR5R[X5X45 UR5R X#5 g)a>Add text data to summary. Args: tag: Data identifier text_string: String to save global_step: Global step value to record walltime: Optional override default walltime (time.time()) of event Examples:: writer.add_text('lstm', 'This is an lstm', 0) writer.add_text('rnn', 'This is an rnn', 10) N)rr>rrlog_text)r'r~rBr=r7s r)add_textSummaryWriter.add_textxs<$ ++  "K ;  ))+Cr,cUR5R[U55 UR5R U5 g)z`Add onnx graph to TensorBoard. Args: onnx_model_file (string): The path to the onnx model. N)rrFr r log_asset)r'onnx_model_files r)rFSummaryWriter.add_onnx_graphs7 ../OP  **?;r,cUR5R[U55 UR5R U5 g)zoAdd openvino graph to TensorBoard. Args: xmlname (string): The path to the openvino model. (the xml file) N)rrr rrH)r'xmlnames r)r SummaryWriter.add_openvino_graphs8 223Fw3OP  **73r,c XSSKJn UR5RU"XX4S95 g)aAdd graph data to summary. The graph is actually processed by `torch.utils.tensorboard.add_graph()` Args: model (torch.nn.Module): Model to draw. input_to_model (torch.Tensor or list of torch.Tensor): A variable or a tuple of variables to be fed. verbose (bool): Whether to print graph structure in console. use_strict_trace (bool): Whether to pass keyword argument `strict` to `torch.jit.trace`. Pass False when you want the tracer to record your mutable container types (list, dict) r)rE)use_strict_traceN)&torch.utils.tensorboard._pytorch_graphrErrB)r'modelinput_to_modelverboserOrEs r)rBSummaryWriter.add_graphs&" A ))%w*rsr,cUnURSSR[S555nURSSR[S555nURSSR[S555nU$)N%z%{:02x}/\)replaceformatord)rawstrretvals r)_encodeSummaryWriter._encodeskY%5%5c#h%?@Y%5%5c#h%?@i&6&6s4y&AB r,mat label_imgcdSSKJn U"U5nUcSn[U5RS5SUR U53n[ R RUR5R5U5n [ R"U 5 Ub-URS[U5:XdS5e[X)US 9 UbYURSURS:XdS 5eURS URS :XdS 5e[!X95 UR"S :XdS5e[%X5 ['X#UR5R5XU5 UbUS3OSn UR)5R+XX:S9 g![a [S5 GNf=f)aAdd embedding projector data to summary. Args: mat: A matrix which each row is the feature vector of the data point metadata (list): A list of labels, each element will be converted to string. label_img: Images correspond to each data point. Each image should be square sized. The amount and the size of the images are limited by the Tensorboard frontend, see limits below. global_step: Global step value to record tag: Name for the embedding Shape: mat: :math:`(N, D)`, where N is number of data and D is feature dimension label_img: :math:`(N, C, H, W)`, where `Height` should be equal to `Width`. Also, :math:`\sqrt{N}*W` must be less than or equal to 8192, so that the generated sprite image can be loaded by the Tensorboard frontend (see `tensorboardX#516 `_ for more). Examples:: import keyword import torch meta = [] while len(meta)<100: meta = meta+keyword.kwlist # get some strings meta = meta[:100] for i, v in enumerate(meta): meta[i] = v+str(i) label_img = torch.rand(100, 3, 32, 32) for i in range(100): label_img[i]*=i/100.0 writer.add_embedding(torch.randn(100, 5), metadata=meta, label_img=label_img) writer.add_embedding(torch.randn(100, 5), label_img=label_img) writer.add_embedding(torch.randn(100, 5), metadata=meta) rrNrrWzKwarning: Embedding dir exists, did you set global_step for add_embedding()?z&#labels should equal with #data points)metadata_headerz&#images should equal with #data pointsz6Image should be square, see tensorflow/tensorboard#670z@mat should be 2D, where mat.size(0) is the number of data pointsz.json)template_filename)rrrhzfillr^rrrrr0makedirsOSErrorr#r-rr r ndimrrr log_embedding) r'r`metadatarar=r~rdrsubdir save_pathrgs r) add_embeddingSummaryWriter.add_embeddingsh #cl  K $**1-.a S0A/BCGGLL!6!6!8!C!C!EvN  _ KK "  99Q<3$ DC D X/ J  99Q<9??1#55 _7_ _5??1%);; u=u u;  -xx1}```} X**,7796PS U-0_se5M$  ..si.m% _ ] _ _s?FF/.F/ predictionsnum_thresholdscSSKJn U"U5U"U5p2[XX5U5n UR5R U XG5 UR 5R XXTS9 g)aAdds precision recall curve. Plotting a precision-recall curve lets you understand your model's performance under different threshold settings. With this function, you provide the ground truth labeling (T/F) and prediction confidence (usually the output of your model) for each target. The TensorBoard UI will let you choose the threshold interactively. Args: tag: Data identifier labels: Ground truth data. Binary label for each element. predictions: The probability that an element be classified as true. Value should in [0, 1] global_step: Global step value to record num_thresholds: Number of thresholds used to draw the curve. walltime: Override default walltime (time.time()) of event Examples:: from tensorboardX import SummaryWriter import numpy as np labels = np.random.randint(2, size=100) # binary label predictions = np.random.rand(100) writer = SummaryWriter() writer.add_pr_curve('pr_curve', labels, predictions, 0) writer.close() rrrN)rrrrr>r log_pr_data) r'r~r+rrr=rsweightsr7rr<s r) add_pr_curveSummaryWriter.add_pr_curvesaJ #%fow{/C 3 WM ++   #  ,,S>,\r,true_positive_countsfalse_positive_countstrue_negative_countsfalse_negative_counts precisionrecallc UR5R[UUUUUUUU U 5 UU 5 UR5R XUUUUUUU U S9 g)aAdds precision recall curve with raw data. Args: tag: Data identifier global_step: Global step value to record num_thresholds (int): Number of thresholds used to draw the curve. walltime: Optional override default walltime (time.time()) of event see: `Tensorboard refenence `_ ) r6ryrzr{r|r}r~rsrvN)rr>rrlog_pr_raw_data) r'r~ryrzr{r|r}r~r=rsrvr7s r)add_pr_curve_rawSummaryWriter.add_pr_curve_rawCs~. ++ -.-."'  "     00FZG\FZG\;D8>@N9@ 1 Br,tagscategorytitlecbX#SU/00nUR5R[U55 g)aShorthand for creating multilinechart. Similar to ``add_custom_scalars()``, but the only necessary argument is *tags*. Args: tags: list of tags that have been used in ``add_scalar()`` Examples:: writer.add_custom_scalars_multilinechart(['twse/0050', 'twse/2330']) MultilineNrr>rr'rrrlayouts r)!add_custom_scalars_multilinechart/SummaryWriter.add_custom_scalars_multilinechartps2[$$789 ++N6,BCr,c[U5S:XdeX#SU/00nUR5R[U55 g)a@Shorthand for creating marginchart. Similar to ``add_custom_scalars()``, but the only necessary argument is *tags*, which should have exactly 3 elements. Args: tags: list of tags that have been used in ``add_scalar()`` Examples:: writer.add_custom_scalars_marginchart(['twse/0050', 'twse/2330', 'twse/2006']) rfMarginN)rrr>rrs r)add_custom_scalars_marginchart,SummaryWriter.add_custom_scalars_marginchartsB4yA~~Xt$456 ++N6,BCr,rcTUR5R[U55 g)aCreate special chart by collecting charts tags in 'scalars'. Note that this function can only be called once for each SummaryWriter() object. Because it only provides metadata to tensorboard, the function can be called before or after the training loop. See ``examples/demo_custom_scalars.py`` for more. Args: layout: {categoryName: *charts*}, where *charts* is also a dictionary {chartName: *ListOfProperties*}. The first element in *ListOfProperties* is the chart's type (one of **Multiline** or **Margin**) and the second element should be a list containing the tags you have used in add_scalar function, which will be collected into the new chart. Examples:: layout = {'Taiwan':{'twse':['Multiline',['twse/0050', 'twse/2330']]}, 'USA':{ 'dow':['Margin', ['dow/aaa', 'dow/bbb', 'dow/ccc']], 'nasdaq':['Margin', ['nasdaq/aaa', 'nasdaq/bbb', 'nasdaq/ccc']]}} writer.add_custom_scalars(layout) Nr)r'rs r)add_custom_scalars SummaryWriter.add_custom_scalarss * ++N6,BCr,verticescolorsfacesc UR5R[XX4U5Xg5 UR5R XX4XVU5 g)aAdd meshes or 3D point clouds to TensorBoard. The visualization is based on Three.js, so it allows users to interact with the rendered object. Besides the basic definitions such as vertices, faces, users can further provide camera parameter, lighting condition, etc. Please check https://threejs.org/docs/index.html#manual/en/introduction/Creating-a-scene for advanced usage. Args: tag: Data identifier vertices: List of the 3D coordinates of vertices. colors: Colors for each vertex faces: Indices of vertices within each triangle. (Optional) config_dict: Dictionary with ThreeJS classes names and configuration. global_step: Global step value to record walltime: Optional override default walltime (time.time()) seconds after epoch of event Shape: vertices: :math:`(B, N, 3)`. (batch, number_of_vertices, channels). If Nothing show on tensorboard, try normalizing the values to [-1, 1]. colors: :math:`(B, N, 3)`. The values should lie in [0, 255]. faces: :math:`(B, N, 3)`. The values should lie in [0, number_of_vertices] for type `uint8`. Expected result after running ``examples/demo_mesh.py``: .. image:: _static/img/tensorboard/add_mesh.png :scale: 30 % N)rr>rrlog_mesh)r'r~rrr config_dictr=r7s r)add_meshSummaryWriter.add_meshsIN ++D{,[]hs  ))#*5H Nr,cURcgURR5H#nUR5 UR5 M% S=UlUlUR 5R 5 SUlg)zClose the current SummaryWriter. This call flushes the unfinished write operation. Use context manager (with statement) whenever it's possible. N)rrrIrLrrendrr'writers r)rLSummaryWriter.closesm    # &&--/F LLN LLN0/324+  $$&!r,cURcgURR5HnUR5 M g)zForce the data in memory to be flushed to disk. Use this call if tensorboard does not update reqularly. Another way is to set the `flush_secs` when creating the SummaryWriter. N)rrrIrs r)rISummaryWriter.flushs5    # &&--/F LLN0r,cU$r$r3r/s r) __enter__SummaryWriter.__enter__s r,c$UR5 gr$)rL)r'exc_typeexc_valexc_tbs r)__exit__SummaryWriter.__exit__s  r,rQc## UR5RXS9 Uv SSS5 g!,(df  g=f7f)a+Context manager to temporarily set default metadata for all enclosed :meth:`add_*` calls. Args: global_step: Global step value to record walltime: Walltime to record (defaults to time.time()) Examples:: with writer.use_metadata(global_step=10): writer.add_scalar("loss", 3.0) rQN)rrTrSs r)rTSummaryWriter.use_metadatas2 " " $ 1 1k 1 ]J^ ] ]s?. ? <?)rrrkrrrrrrrrr(rrrV)NNrr)NrNN)NNr)NNr3)NNrN)NTN)NNNTCHW)NiDN)NFT)NNNdefaultN)NNN)runtitled)NNNNN)3rWrXrYrZr[rrhrvboolrr*rrrrrrnumpy_compatiblerrrrrrr(r!r0r5r:r@rErFrrB staticmethodr^rprwrrrrrrLrIrrr\r]rTr^r3r,r)rrsK%)%'(,')(+-/,0%),6+=]$SM]$c]]$! ]$  } ]$ ! ]$&c]]$$D>]$c]]$#4.]$~ N *"#')- .Oc5sE3)>#??@.Oc5j).O3- .O "# .Oh*.(,*,13$Z$Z '7 78$Z"# $Z uo $Z #3- $Z"*#$ZT*.(, 0Z0Z"#u*-0Z"# 0Z uo 0Zd *.".(,'I'I%'I"# 'I 3- 'I uo 'If*.(,@S@S"#@Suo@SL*.(,). 8`8`)8`"# 8` uo 8` "# 8`|*.(,)/ =`=`)=`"# =` uo =` "# =`H*.(,).*.&`&`)&`) &` "# &` uo &`"#&`T#Y'&`X*.$((, jj"# j D> j uo j6*./0(,)0``)`"# ` %U +, ` uo `"#`B*.).(, [[)["# [ "# [ uo [8*.(, DDD"# D uo D, < 4 ! t(*.)- Qn!Qn( Qn "# Qnp*.,/(,-]-]%-]* -] "# -] %SM -]uo-]p*.,/(,+B+B#3+B$4 +B #3 +B $4 +B(+B%+B"#+B%SM+Buo+B`&# Ds)DD D*&# Ds)DD D&Dd39o-.D6(,&*)-(,)N)N')N% )N $ )N"#)Nuo)NV "*.  r,r)9r[rir\rloggingrrttypingrrr& comet_utilsr embeddingrrr r event_file_writerr onnx_graphr openvino_graphr protorrproto.event_pb2rrr<rrrrrrrrrrrrrutilsr getLoggerrWr.ndarrayrr$Tensor ImportErrorr!r`rr3r,r)rs.  " $DD.'/).#   8 $== || ""Jd$d$NDDa  sC CC