i#dZddlZddlZddlmZddlmZddlmZddl Z ddl m cm Z ddl m Z ddlmZdd lmZdd lmZdd lmZmZdd lmZmZdd lmZddlmZddlm Z m!Z!m"Z"m#Z#m$Z$m%Z%ddl&m'Z'ddl(m)Z)ddl*m+Z+m,Z,m-Z-e$j\e/Z0dZ1dZ2dZ3de jhde jhfdZ5ee"dGdde Z6ee"dGdd e Z7ee"Gd!d"e Z8Gd#d$e jrZ:Gd%d&e jrZ;Gd'd(e jrZ<Gd)d*e jrZ=Gd+d,e jrZ>Gd-d.e jrZ?Gd/d0e jrZ@Gd1d2e jrZAGd3d4e jrZBGd5d6eZCGd7d8e jrZDGd9d:e jrZEGd;de jrZG ded?e jrd@e jhdAe jhdBe jhdCe jhdzdDeHdEeHfdFZIGdGdHe jrZJGdIdJe jrZKGdKdLe jrZLGdMdNe jrZMGdOdPe jrZNGdQdReZOGdSdTe jrZPGdUdVe jrZQe"GdWdXeZRGdYdZeRZSe"d[Gd\d]eRZTe"Gd^d_eRZUe"Gd`daeRZVe"GdbdceRZWgddZXy)fzPyTorch CLAP model.N)Callable) dataclass)Any)nn)initialization)ACT2FN)GradientCheckpointingLayer)BaseModelOutputBaseModelOutputWithPooling)ALL_ATTENTION_FUNCTIONSPreTrainedModel)Unpack)apply_chunking_to_forward) ModelOutputTransformersKwargsauto_docstringcan_return_tuplelogging torch_int)merge_with_config_defaults)capture_outputs)ClapAudioConfig ClapConfigClapTextConfigc|j\}}}|dddddddfjdd|d}|j|||z|}|S)ae Interpolate data in time domain. This is used to compensate the resolution reduction in downsampling of a CNN. Args: hidden_states (`torch.FloatTensor` of shape (batch_size, time_length, classes_num)): Input hidden states ratio (`int`): The ratio of the length of the output to the length of the input. Nr)shaperepeatreshape) hidden_statesratio batch_size time_length classes_num upsampleds w/var/www/vps2.regionflexible.com/Desarrollo/venv/lib/python3.12/site-packages/transformers/models/clap/modeling_clap.py interpolater(.sX.;-@-@*ZkaD!m,33Aq%CI!!*kE.A;OI c|j\}}}}|j|||z|||z||}|jddddddjjd|||}|S)aR Returns the resized hidden states. The output shape should be `(batch_size * num_windows, window_size, window_size, num_channels)` Args: hidden_states (`torch.FloatTensor` of shape `(batch_size, height, width, num_channels)`): Input hidden states window_size (`int`): Window size rrrrviewpermute contiguous)r! window_sizer#heightwidth num_channelswindowss r'window_partitionr8?s}/<.A.A+J|!&&Fk);8Lk[gM##Aq!Q15@@BGGKYdfrsG Nr)c|jd}|jd||z||z|||}|jddddddjjd|||}|S)a Merges windows to produce higher resolution features. Args: windows (`torch.FloatTensor` of shape `(num_windows * batch_size, window_size, window_size, num_channels)`): Input windows window_size (`int`): Window size height (`int`): Height of the resized audio width (`int`): Width of the resized audio r.rrrr+r,r-r/)r7r3r4r5r6s r'window_reverser:Tsn==$Lll2v4e{6JKYdfrsGooaAq!Q/::<AA"feUabG Nr)logitsreturnctjt||j}tj j ||S)Ndevice)torcharangelenr?r functional cross_entropy)r;labelss r'contrastive_lossrFis1 \\#f+fmm