Train.py

import sys
import random
import torch
import copy

from Transition import Transition, getMissingLinks, applyTransition
import Features
from Dicts import Dicts
from Util import timeStamp, prettyInt, numParameters, getDevice
from Rl import ReplayMemory, selectAction, optimizeModel
import Networks
import Decode
import Config

from conll18_ud_eval import load_conllu, evaluate

################################################################################
def trainMode(debug, filename, type, modelDir, nbIter, batchSize, devFile, bootstrapInterval, silent=False) :
  transitionSet = [Transition(elem) for elem in ["RIGHT", "LEFT", "SHIFT", "REDUCE"]]
  strategy = {"RIGHT" : 1, "SHIFT" : 1, "LEFT" : 0, "REDUCE" : 0}

  sentences = Config.readConllu(filename)

  if type == "oracle" :
    trainModelOracle(debug, modelDir, filename, nbIter, batchSize, devFile, transitionSet, strategy, sentences, bootstrapInterval, silent)
    return

  if type == "rl":
    trainModelRl(debug, modelDir, filename, nbIter, batchSize, devFile, transitionSet, strategy, sentences, silent)
    return

  print("ERROR : unknown type '%s'"%type, file=sys.stderr)
  exit(1)
################################################################################

################################################################################
def extractExamples(debug, ts, strat, config, dicts, network, dynamic) :
  examples = []
  with torch.no_grad() :
    EOS = Transition("EOS")
    config.moveWordIndex(0)
    moved = True
    while moved :
      missingLinks = getMissingLinks(config)
      candidates = sorted([[trans.getOracleScore(config, missingLinks), trans] for trans in ts if trans.appliable(config)])
      if len(candidates) == 0 :
        break
      best = min([cand[0] for cand in candidates])
      candidateOracle = random.sample([cand for cand in candidates if cand[0] == best], 1)[0][1]
      features = network.extractFeatures(dicts, config)
      candidate = candidateOracle.name
      if debug :
        config.printForDebug(sys.stderr)
        print(str([[c[0],c[1].name] for c in candidates])+"\n"+("-"*80)+"\n", file=sys.stderr)
      if dynamic :
        output = network(features.unsqueeze(0).to(getDevice()))
        scores = sorted([[float(output[0][index]), ts[index].appliable(config), ts[index].name] for index in range(len(ts))])[::-1]
        candidate = [[cand[0],cand[2]] for cand in scores if cand[1]][0][1]
        if debug :
          print(candidate.name, file=sys.stderr)
        
      goldIndex = [trans.name for trans in ts].index(candidateOracle.name)
      candidateIndex = [trans.name for trans in ts].index(candidate)
      example = torch.cat([torch.LongTensor([goldIndex]), features])
      examples.append(example)

      moved = applyTransition(ts, strat, config, candidate)
  
    EOS.apply(config)
  
  return examples
################################################################################

################################################################################
def evalModelAndSave(debug, model, dicts, modelDir, devFile, bestLoss, totalLoss, bestScore, epoch, nbIter) :
  devScore = ""
  saved = True if bestLoss is None else totalLoss < bestLoss
  bestLoss = totalLoss if bestLoss is None else min(bestLoss, totalLoss)
  if devFile is not None :
    outFilename = modelDir+"/predicted_dev.conllu"
    Decode.decodeMode(debug, devFile, "model", modelDir, model, dicts, open(outFilename, "w"))
    res = evaluate(load_conllu(open(devFile, "r")), load_conllu(open(outFilename, "r")), [])
    UAS = res["UAS"][0].f1
    score = UAS
    saved = True if bestScore is None else score > bestScore
    bestScore = score if bestScore is None else max(bestScore, score)
    devScore = ", Dev : UAS=%.2f"%(UAS)
  if saved :
    torch.save(model, modelDir+"/network.pt")
  print("{} : Epoch {:{}}/{}, loss={:6.2f}{} {}".format(timeStamp(), epoch, len(str(nbIter)), nbIter, totalLoss, devScore, "SAVED" if saved else ""), file=sys.stderr)

  return bestLoss, bestScore
################################################################################

################################################################################
def trainModelOracle(debug, modelDir, filename, nbEpochs, batchSize, devFile, transitionSet, strategy, sentencesOriginal, bootstrapInterval, silent=False) :
  dicts = Dicts()
  dicts.readConllu(filename, ["FORM","UPOS"], 2)
  dicts.save(modelDir+"/dicts.json")
  network = Networks.BaseNet(dicts, len(transitionSet)).to(getDevice())
  examples = []
  sentences = copy.deepcopy(sentencesOriginal)
  print("%s : Starting to extract examples..."%(timeStamp()), file=sys.stderr)
  for config in sentences :
    examples += extractExamples(debug, transitionSet, strategy, config, dicts, network, False)
  print("%s : Extracted %s examples"%(timeStamp(), prettyInt(len(examples), 3)), file=sys.stderr)
  examples = torch.stack(examples)

  print("%s : Model has %s parameters"%(timeStamp(), prettyInt((numParameters(network)), 3)), file=sys.stderr)
  optimizer = torch.optim.Adam(network.parameters(), lr=0.0001)
  lossFct = torch.nn.CrossEntropyLoss()
  bestLoss = None
  bestScore = None
  for epoch in range(1,nbEpochs+1) :
    if bootstrapInterval is not None and epoch > 1 and (epoch-1) % bootstrapInterval == 0 :
      examples = []
      sentences = copy.deepcopy(sentencesOriginal)
      print("%s : Starting to extract dynamic examples..."%(timeStamp()), file=sys.stderr)
      for config in sentences :
        examples += extractExamples(debug, transitionSet, strategy, config, dicts, network, True)
      print("%s : Extracted %s examples"%(timeStamp(), prettyInt(len(examples), 3)), file=sys.stderr)
      examples = torch.stack(examples)

    network.train()
    examples = examples.index_select(0, torch.randperm(examples.size(0)))
    totalLoss = 0.0
    nbEx = 0
    printInterval = 2000
    advancement = 0
    for batchIndex in range(0,examples.size(0)-batchSize,batchSize) :
      batch = examples[batchIndex:batchIndex+batchSize].to(getDevice())
      targets = batch[:,:1].view(-1)
      inputs = batch[:,1:]
      nbEx += targets.size(0)
      advancement += targets.size(0)
      if not silent and advancement >= printInterval :
        advancement = 0
        print("Curent epoch %6.2f%%"%(100.0*nbEx/examples.size(0)), end="\r", file=sys.stderr)
      outputs = network(inputs)
      loss = lossFct(outputs, targets)
      network.zero_grad()
      loss.backward()
      optimizer.step()
      totalLoss += float(loss)

    bestLoss, bestScore = evalModelAndSave(debug, network, dicts, modelDir, devFile, bestLoss, totalLoss, bestScore, epoch, nbEpochs)
################################################################################

################################################################################
def trainModelRl(debug, modelDir, filename, nbIter, batchSize, devFile, transitionSet, strategy, sentencesOriginal, silent=False) :

  memory = None
  dicts = Dicts()
  dicts.readConllu(filename, ["FORM", "UPOS"])
  dicts.save(modelDir + "/dicts.json")

  policy_net = Networks.BaseNet(dicts, len(transitionSet)).to(getDevice())
  target_net = Networks.BaseNet(dicts, len(transitionSet)).to(getDevice())
  target_net.load_state_dict(policy_net.state_dict())
  target_net.eval()
  policy_net.train()
  optimizer = torch.optim.Adam(policy_net.parameters(), lr=0.0001)
  print("%s : Model has %s parameters"%(timeStamp(), prettyInt((numParameters(policy_net)), 3)), file=sys.stderr)

  bestLoss = None
  bestScore = None

  sentences = copy.deepcopy(sentencesOriginal)
  nbExByEpoch = sum(map(len,sentences))
  sentIndex = 0

  for epoch in range(1,nbIter+1) :
    i = 0
    totalLoss = 0.0
    while True :
      if sentIndex >= len(sentences) :
        sentences = copy.deepcopy(sentencesOriginal)
        random.shuffle(sentences)
        sentIndex = 0

      if not silent :
        print("Curent epoch %6.2f%%"%(100.0*i/nbExByEpoch), end="\r", file=sys.stderr)
      sentence = sentences[sentIndex]
      sentence.moveWordIndex(0)
      state = policy_net.extractFeatures(dicts, sentence).to(getDevice())

      while True :
        missingLinks = getMissingLinks(sentence)
        if debug :
          sentence.printForDebug(sys.stderr)
        action = selectAction(policy_net, state, transitionSet, sentence, missingLinks, probaRandom=0.1, probaOracle=0.1)
        if action is None :
          break

        appliable = action.appliable(sentence)

        # Reward for doing an illegal action
        reward = -3.0
        if appliable :
          reward = -1.0*action.getOracleScore(sentence, missingLinks)
      
        reward = torch.FloatTensor([reward]).to(getDevice())

        newState = None
        if appliable :
          applyTransition(transitionSet, strategy, sentence, action.name)
          newState = policy_net.extractFeatures(dicts, sentence).to(getDevice())

        if memory is None :
          memory = ReplayMemory(5000, state.numel())
        memory.push(state, torch.LongTensor([transitionSet.index(action)]).to(getDevice()), newState, reward)
        state = newState
        if i % batchSize == 0 :
          totalLoss += optimizeModel(batchSize, policy_net, target_net, memory, optimizer)
          if i % (2*batchSize) == 0 :
            target_net.load_state_dict(policy_net.state_dict())
            target_net.eval()
            policy_net.train()
        i += 1

        if state is None :
          break
      if i >= nbExByEpoch :
        break
      sentIndex += 1
    bestLoss, bestScore = evalModelAndSave(debug, policy_net, dicts, modelDir, devFile, bestLoss, totalLoss, bestScore, epoch, nbIter)
################################################################################