Untitled
unknown
python
2 years ago
20 kB
18
Indexable
# %%
import argparse
import gzip
import os
import pickle
import time
from copy import deepcopy
from models import MLP
import numpy as np
import pandas as pd
import soundfile as sf
import torch
from torch.nn.utils.rnn import pad_sequence
from tqdm import tqdm
import csv
import data
import models
from collections import defaultdict
from utils import get_audio_duration
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
config = data.read_config("experiments/no_unfreezing.cfg")
train_dataset, valid_dataset, test_dataset = data.get_SLU_datasets(config)
def create_csv_file(filename, headers):
with open(filename, 'w', newline='') as file:
writer = csv.writer(file)
writer.writerow(headers)
def write_to_csv(filename, data):
with open(filename, 'a', newline='') as file:
writer = csv.writer(file)
writer.writerow(data)
bucket_results_file = '../slurp_multicache_audio_per_bucket_dets.csv'
if not os.path.exists(bucket_results_file):
headers = ['SpeakerId', 'l1_threshold', 'l2_threshold', 'bucket', 'train', 'test', 'l1 tp', 'l1 hits',
'l1_hit_rate', 'l1_cache_acc', 'l2_sample', 'l2_tp', 'l2_hits', 'l2_hit_rate', 'l2_cache_acc']
create_csv_file(bucket_results_file, headers)
# results_file = 'slurp_multicache_audio_bucket.csv'
# if not os.path.exists(results_file):
# headers = ['SpeakerId', 'l1_threshold', 'l2_threshold', 'train', 'test', 'l1_hit_rate', 'l1_cache_acc', 'l2_hit_rate', 'l2_cache_acc', 'total_acc']
# create_csv_file(results_file, headers)
L1_THRESHOLDS = [400, 700, 1100]
# L2_THRESHOLDS = [30, 85, 170]
# L2_THRESHOLDS = [50, 110, 200]
L2_THRESHOLDS = [25, 50, 100]
# L2_THRESHOLDS = [30,60,110]
# L1_THRESHOLDS = [400, 600, 1100]
# L2_THRESHOLDS = [40, 90, 200]
# L2_THRESHOLDS = [30, 85, 170]
# variables to save #hits, #corrects
cumulative_l1_hits, cumulative_l1_corrects, cumulative_l1_hit_correct = 0, 0, 0
cumulative_l2_hits, cumulative_l2_corrects, cumulative_l2_hit_correct = 0, 0, 0
buckets = [1, 2, 3]
parser = argparse.ArgumentParser(description="input cutoff")
parser.add_argument("--cutoff", type=float, default=2.0)
parser.add_argument("--dynamic", type=bool, default=False)
args = parser.parse_args()
cutoff = args.cutoff
dynamic = args.dynamic
pwd = os.getcwd()
wav_path = os.path.join(pwd, 'SLURP/slurp_real/')
folder_path = os.path.join(pwd, f'models/SLURP/slurp_multicache_selective_audio_bucket_k_{cutoff}')
# folder_path = os.path.join(pwd, 'models/SLURP/curated-slurp-headset')
# folder_path = os.path.join(pwd, 'models/SLURP/curated-slurp-without-headset')
# folder_path = os.path.join(pwd, 'models/SLURP/curated-slurp-headset-base')
# slurp-pretrained.pth/in-domain
# pretrained_file = "slurp-pretrained.pth"
# pretrained_path = os.path.join(pwd + "/models/SLURP/", pretrained_file)
# base path
pretrained_path = "experiments/no_unfreezing/training/model_state.pth"
cloud_model = models.Model(config).eval()
cloud_model.load_state_dict(
torch.load(pretrained_path, map_location=device)) # load trained model
def multicache_test(model, df, cluster_ids, cluster_centers, transcript_list, training_idxs, intent_list, L1_THRESHOLD,
L2_THRESHOLD):
# ----------------- prepare for cluster -----------------
cluster_id_length = torch.tensor(list(map(len, cluster_ids)), dtype=torch.long, device=device)
cluster_ids = pad_sequence(cluster_ids, batch_first=True, padding_value=0).to(device)
cluster_centers = torch.stack(cluster_centers).to(device)
# ----------------- prepare for phoneme -----------------
# prepare all the potential phoneme sequences
label_lengths = torch.tensor(list(map(len, phoneme_list)), dtype=torch.long)
phoneme_label = pad_sequence(phoneme_list, batch_first=True).to(device)
# no reduction, loss on every sequence
ctc_loss_k_means_eval = torch.nn.CTCLoss(reduction='none')
ctc_loss_phoneme_eval = torch.nn.CTCLoss(reduction='none')
# ------------------ variables to record performance --------------------
tp, total, hits, l1_hits, l2_hits, l1_correct, l2_correct, l2_total, l1_total = 0, 0, 0, 0, 0, 0, 0, 0, 0
for _, row in df.iterrows():
if row[0] in training_idxs:
continue
# print(row['sentence'])
# # of total evaluation samples
total += 1
wav = os.path.join(wav_path, row['recording_path'])
x, _ = sf.read(wav)
x = torch.tensor(x, dtype=torch.float, device=device).unsqueeze(0)
with torch.no_grad():
if bucket == 1:
# print('just do L2')
l2_total += 1
phoneme_pred = model.pretrained_model.compute_phonemes(x)
# repeat it #sentence times to compare with ground truth
phoneme_pred = phoneme_pred.repeat(1, phoneme_label.shape[0], 1)
pred_lengths = torch.full(size=(phoneme_label.shape[0],), fill_value=phoneme_pred.shape[0],
dtype=torch.long)
loss = ctc_loss_phoneme_eval(phoneme_pred, phoneme_label, pred_lengths, label_lengths)
# print('l2 loss ', loss.min())
# loss = torch.nan_to_num(loss, nan=float('inf')) # remove potential nan from loss
pred_result = loss.argmin()
if torch.isnan(loss).any():
print('nan eval on speaker: %s' % user_id)
if dynamic:
# filename = 'utils/slurp-MLP-L2.pkl'
filename = 'utils/slurp-MLP-L2_old.pkl'
with open(filename, 'rb') as f:
load_data = pickle.load(f)
MLP_model = load_data['model']
dur = torch.tensor([[get_audio_duration(wav)]], dtype=torch.float32)
L2_THRESHOLD = MLP_model(dur).item()
if loss.min() <= L2_THRESHOLD:
# print('l2 hit: ', row['sentence'])
l2_hits += 1
if row['intent'] == intent_list[pred_result]:
l2_correct += 1
else:
# ----------------- l1 -------------------
x_feature = cloud_model.pretrained_model.compute_cnn_features(x)
dists = torch.cdist(x_feature, cluster_centers)
dists = dists.max(dim=-1)[0].unsqueeze(-1) - dists
pred = dists.swapaxes(1, 0)
pred_lengths = torch.full(size=(cluster_ids.shape[0],), fill_value=pred.shape[0], dtype=torch.long)
loss = ctc_loss_k_means_eval(pred.log_softmax(dim=-1), cluster_ids, pred_lengths, cluster_id_length)
pred_intent = loss.argmin().item()
# print('l1 loss ', loss[pred_intent])
l1_total += 1
if dynamic:
filename = 'utils/k_means_MLP.pkl'
with open(filename, 'rb') as f:
load_data = pickle.load(f)
MLP_model_k_means = load_data['model']
dur = torch.tensor([[get_audio_duration(wav)]], dtype=torch.float32)
L1_THRESHOLD = MLP_model_k_means(dur).item()
print(round(dur.item(), 4), L1_THRESHOLD)
if loss[pred_intent] < L1_THRESHOLD:
# go with l1: kmeans
# print('l1 hit: ', row['sentence'])
l1_hits += 1
if row['intent'] == intent_list[pred_intent]:
l1_correct += 1
else:
# ------------------ l2 -------------------
# phoneme_pred = model.compute_phoneme_from_features(x_feature) #doesnt work RuntimeError: input must have 3 dimensions, got 5
l2_total += 1
phoneme_pred = model.pretrained_model.compute_phonemes(x)
# repeat it #sentence times to compare with ground truth
phoneme_pred = phoneme_pred.repeat(1, phoneme_label.shape[0], 1)
pred_lengths = torch.full(size=(phoneme_label.shape[0],), fill_value=phoneme_pred.shape[0],
dtype=torch.long)
loss = ctc_loss_phoneme_eval(phoneme_pred, phoneme_label, pred_lengths, label_lengths)
# print('l2 loss ', loss.min())
# loss = torch.nan_to_num(loss, nan=float('inf')) # remove potential nan from loss
pred_result = loss.argmin()
if torch.isnan(loss).any():
print('nan eval on speaker: %s' % user_id)
if dynamic:
# filename = 'utils/slurp-MLP-L2.pkl'
filename = 'utils/slurp-MLP-L2_old.pkl'
with open(filename, 'rb') as f:
load_data = pickle.load(f)
MLP_model = load_data['model']
dur = torch.tensor([[get_audio_duration(wav)]], dtype=torch.float32)
L2_THRESHOLD = MLP_model(dur).item()
if loss.min() <= L2_THRESHOLD:
# print('l2 hit: ', row['sentence'])
l2_hits += 1
if row['intent'] == intent_list[pred_result]:
l2_correct += 1
# else:
# print('%s,%s' % (row['sentence'], transcript_list[pred_result]))
# else:
# # do the calculation
# # cloud_model.predict_intents(x)
# # print('cloud. loss was %f ' % loss.min())
#
return total, l1_total, l2_total, l1_hits, l2_hits, l1_correct, l2_correct
slurp_df = pd.read_csv('SLURP/csv/slurp_mini_FE_MO_ME_FO_UNK.csv')
# slurp_df = pd.read_csv('SLURP/csv/slurp_headset.csv')
# slurp_df = pd.read_csv('SLURP/csv/slurp_without_headset.csv')
slurp_df = deepcopy(slurp_df)
num_nan_train, nan_nan_eval = 0, 0
speakers = np.unique(slurp_df['user_id'])
# starts_with = 'UNK'
# speakers = [elem for elem in speakers if not elem.startswith(starts_with)]
# speakers = ['MO-433', 'UNK-326', 'FO-232', 'ME-144']
# speakers = ['FO-234', 'FO-462', 'FO-488', 'FO-493', 'ME-144', 'ME-473']
# speakers = ['FO-234', 'FO-462', 'FO-488', 'ME-144', 'ME-369', 'ME-473','MO-030', 'MO-038']
# speakers = ['FO-234']
cumulative_l1_sample, cumulative_l2_sample, cumulative_correct, cumulative_hit, cumulative_cache_miss, cumulative_hit_incorrect, cumulative_hit_rate, cumulative_cache_acc, cumulative_acc, total_train = 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
cumulative_sample, cumulative_hit_correct, cumulative_hits = 0, 0, 0
bckt = defaultdict(lambda: defaultdict(lambda: defaultdict(int)))
for _, user_id in tqdm(enumerate(speakers), total=len(speakers)):
# print(f'EVAL FOR SPEAKER {user_id} cutoff {cutoff}')
print(f'EVAL FOR SPEAKER {user_id}')
print('count ', slurp_df[slurp_df['user_id'] == user_id].shape[0])
bckt_sample, l1_bckt_sample, l2_bckt_sample, l1_bckt_correct, l2_bckt_correct, l1_bckt_hit, l2_bckt_hit, l1_bckt_hit_correct, l2_bckt_hit_correct, bckt_train = 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
for bucket in buckets:
filename = f'slurp_model_multicache_{user_id}_audio_bucket_{bucket}_cutoff_{cutoff}'
# filename = f'slurp_curated_multicache_{user_id}_audio_bucket_{bucket}'
# filename = f'slurp_curated_wo_headset_multicache_{user_id}_audio_bucket_{bucket}'
# filename = f'slurp_curated_headset_base_multicache_{user_id}_audio_bucket_{bucket}'
file_path = os.path.join(folder_path, filename + '.pth')
model = models.Model(config)
model.load_state_dict(torch.load(file_path, map_location=device))
with gzip.open(os.path.join(folder_path, filename + '.pkl.gz'), 'rb') as f:
metadata = pickle.load(f)
user_id = metadata['speakerId']
df = metadata['df']
transcript_list = metadata['transcript_list']
phoneme_list = metadata['phoneme_list']
intent_list = metadata['intent_list']
training_idxs = metadata['training_idxs']
cluster_ids = metadata['cluster_ids']
cluster_centers = metadata['cluster_centers']
print(f'bucket {bucket} train {len(training_idxs)} test: {len(df)-len(training_idxs)}')
total, l1_total, l2_total, l1_hits, l2_hits, l1_correct, l2_correct = 0, 0, 0, 0, 0, 0, 0
if cluster_ids and cluster_centers and phoneme_list:
total, l1_total, l2_total, l1_hits, l2_hits, l1_correct, l2_correct = multicache_test(model, df, cluster_ids,
cluster_centers,
transcript_list, training_idxs,
intent_list,
L1_THRESHOLD=L1_THRESHOLDS[
bucket - 1],
L2_THRESHOLD=L2_THRESHOLDS[
bucket - 1])
bckt[bucket]['l1']['hits'] += l1_hits
bckt[bucket]['l1']['tp'] += l1_correct
bckt[bucket]['l2']['hits'] += l2_hits
bckt[bucket]['l2']['tp'] += l2_correct
bckt[bucket]['l1']['total'] += l1_total
bckt[bucket]['l2']['total'] += l2_total
l1_hit_rate, l1_cache_acc, l2_hit_rate, l2_cache_acc = 0, 0, 0, 0
bckt_train += len(training_idxs)
bckt_sample += total
l1_bckt_sample += l1_total
l2_bckt_sample += l2_total
if total >= 5: # skip for users with < 5 eval samples
# print('EVAL FOR SPEAKER %s: BUCKET %d' % (user_id, bucket))
# l1_bckt_correct += l1_correct + (total - l1_hits)
l1_bckt_hit += l1_hits
l1_bckt_hit_correct += l1_correct
# l2_bckt_correct += l2_correct + (l2_total - l2_hits)
l2_bckt_hit += l2_hits
l2_bckt_hit_correct += l2_correct
if l1_hits:
l1_hit_rate = round((l1_hits / l1_total), 4)
l1_cache_acc = round((l1_correct / l1_hits), 4)
# print('l1_hit_rate ', l1_hit_rate)
# print('l1_cache_acc ', l1_cache_acc)
else:
print('no hits in l1')
if l2_hits:
l2_hit_rate = round((l2_hits / l2_total), 4)
l2_cache_acc = round((l2_correct / l2_hits), 4)
# print('l2_hit_rate ', l2_hit_rate)
# print('l2_cache_acc ', l2_cache_acc)
else:
print('no hits in l2')
print(f'Bucket {bucket} [l1: hit_rate {l1_hit_rate} cache_acc {l1_cache_acc}], [l2: hit_rate {l2_hit_rate} cache_acc {l2_cache_acc}]')
else:
print('not enough samples: %s' % user_id)
# values = [user_id, L1_THRESHOLDS[bucket - 1], L2_THRESHOLDS[bucket - 1], bucket, len(training_idxs), total,
# l1_correct, l1_hits, l1_hit_rate, l1_cache_acc, l2_total, l2_correct, l2_hits, l2_hit_rate,
# l2_cache_acc]
# write_to_csv(bucket_results_file, values)
# print(values)
total_acc, l1_hit_rate, l2_hit_rate, l1_cache_acc, l2_cache_acc = 0, 0, 0, 0, 0
total_train += bckt_train
# following needed for overall evaluation
cumulative_sample += bckt_sample
cumulative_l1_sample += l1_bckt_sample
cumulative_l2_sample += l2_bckt_sample
cumulative_l1_hits += l1_bckt_hit
cumulative_l2_hits += l2_bckt_hit
cumulative_l1_hit_correct += l1_bckt_hit_correct
cumulative_l2_hit_correct += l2_bckt_hit_correct
# cumulative_l1_corrects += l1_bckt_correct #tp+total-hits
# cumulative_l2_corrects += l2_bckt_correct
print(f"------------------------------------{user_id}------------------------------------")
print(f'total train {bckt_train} test {bckt_sample}')
if l1_bckt_sample:
l1_hit_rate = round((l1_bckt_hit / l1_bckt_sample), 4)
# print('l1 hit_rate: %.4f' % l1_hit_rate)
if l1_bckt_hit:
l1_cache_acc = round((l1_bckt_hit_correct / l1_bckt_hit), 4)
# print('l1 cache_acc: %.4f' % l1_cache_acc)
if l2_bckt_sample:
l2_hit_rate = round((l2_bckt_hit / l2_bckt_sample), 4)
# print('l2 hit_rate: %.4f' % l2_hit_rate)
if l2_bckt_hit:
l2_cache_acc = round((l2_bckt_hit_correct / l2_bckt_hit), 4)
# print('l2 cache_acc: %.4f' % l2_cache_acc)
print(f'{user_id} \nl1: hit_rate {l1_hit_rate} cache_acc {l1_cache_acc}, l2: hit_rate {l2_hit_rate} cache_acc {l2_cache_acc}')
if bckt_sample:
total_acc = round((l1_bckt_hit_correct + l2_bckt_hit_correct + (
bckt_sample - l1_bckt_hit - l2_bckt_hit) * 0.9014) / bckt_sample, 4)
# tp+total-hit/total
print('total_acc: %.4f' % total_acc)
print(f"------------------------------------------------------------------------")
values = [user_id, L1_THRESHOLDS, L2_THRESHOLDS, bckt_train, bckt_sample, l1_hit_rate, l1_cache_acc, l2_hit_rate,
l2_cache_acc, total_acc]
# write_to_csv(results_file, values)
# print(values)
# print(bckt)
print(f"------------------------------------------------------------------------")
cumulative_hits = cumulative_l1_hits + cumulative_l2_hits
cumulative_hit_correct = cumulative_l1_hit_correct + cumulative_l2_hit_correct
cumulative_l1_hit_rate, cumulative_l1_hit_acc, cumulative_l2_hit_rate, cumulative_l2_hit_acc = 0, 0, 0, 0
if dynamic:
print('DYNAMIC')
else:
print(L1_THRESHOLDS, L2_THRESHOLDS)
if cumulative_l1_sample:
cumulative_l1_hit_rate = round(cumulative_l1_hits / cumulative_l1_sample, 4)
if cumulative_l1_hits:
cumulative_l1_hit_acc = round(cumulative_l1_hit_correct / cumulative_l1_hits, 4)
if cumulative_l2_sample:
cumulative_l2_hit_rate = round(cumulative_l2_hits / cumulative_l2_sample, 4)
if cumulative_l2_hits:
cumulative_l2_hit_acc = round(cumulative_l2_hit_correct / cumulative_l2_hits, 4)
cumulative_hit_rate = round(cumulative_hits / cumulative_sample, 4)
cumulative_cache_acc = round(cumulative_hit_correct / cumulative_hits, 4)
cloud = (cumulative_sample - cumulative_l1_hits - cumulative_l2_hits) * float(0.9014)
cumulative_acc = round((cumulative_l1_hit_correct + cumulative_l2_hit_correct + cloud) / cumulative_sample, 4)
print(f'total train {total_train} test {cumulative_sample}')
print(f'l1: hit_rate {cumulative_l1_hit_rate} hit_acc {cumulative_l1_hit_acc}')
print(f'l2: hit_rate {cumulative_l2_hit_rate} hit_acc {cumulative_l2_hit_acc}')
print(f'l1+L2: hit_rate {cumulative_hit_rate} hit_acc {cumulative_cache_acc}')
# print(f'acc {cumulative_acc} for cutoff {cutoff}')
print(f'acc {cumulative_acc}')
print(f"------------------------------------------------------------------------")
print('cloud ', (cumulative_sample - cumulative_l1_hits - cumulative_l2_hits))
#
# for i in buckets:
# print('bucket', i)
# if bckt[i]['l1']['total']:
# print(round(bckt[i]['l1']['hits'] / bckt[i]['l1']['total'], 2))
# if bckt[i]['l1']['hits']:
# print(round(bckt[i]['l1']['tp'] / bckt[i]['l1']['hits'], 2))
# if bckt[i]['l2']['total']:
# print(round(bckt[i]['l2']['hits'] / bckt[i]['l2']['total'], 2))
# if bckt[i]['l2']['hits']:
# print(round(bckt[i]['l2']['tp'] / bckt[i]['l2']['hits'], 2))
Editor is loading...