Trie file version mismatch (4 instead of expected 3). Update your trie file

Hi Deepspeech,

  • OS Platform and Distribution (e.g., Linux Ubuntu 16.04) : Ubuntu 16.04 LTS
  • TensorFlow installed from (our builds, or upstream TensorFlow) : NO
  • TensorFlow version (use command below) : 1.13
  • Python version : Python 3.6.8

from the “pip list” command I can see the below installed version of both. Could you please correct me.
deepspeech 0.6.0a4
ds-ctcdecoder 0.5.1

I am getting --> Error: Trie file version mismatch (4 instead of expected 3). Update your trie file.

Here is the output of the command

I Saved new best validating model with loss 149.557864 to: /home/jargon-javi/speech/tools/backup/checkout/best_dev-120
Epoch 1 | Training | Elapsed Time: 0:00:06 | Steps: 5 | Loss: 175.983118
Epoch 1 | Validation | Elapsed Time: 0:00:00 | Steps: 3 | Loss: 150.642143 | Dataset: /home/jargon-javi/speech/tools/backup/train/dev.csv
Epoch 2 | Training | Elapsed Time: 0:00:06 | Steps: 5 | Loss: 174.152228
Epoch 2 | Validation | Elapsed Time: 0:00:00 | Steps: 3 | Loss: 150.533025 | Dataset: /home/jargon-javi/speech/tools/backup/train/dev.csv
Epoch 3 | Training | Elapsed Time: 0:00:06 | Steps: 5 | Loss: 172.228995
Epoch 3 | Validation | Elapsed Time: 0:00:00 | Steps: 3 | Loss: 151.794245 | Dataset: /home/jargon-javi/speech/tools/backup/train/dev.csv
I Early stop triggered as (for last 4 steps) validation loss: 151.794245 with standard deviation: 0.487455 and mean: 150.244344
I FINISHED optimization in 0:00:26.650283
Error: Trie file version mismatch (4 instead of expected 3). Update your trie file.
terminate called after throwing an instance of ‘int’
Aborted (core dumped)

Ok, can you explain what you are doing ? Because this already looks like your issue, mixing incompatible versions.

Hi @lissyx,

I am running ./Deepspeech.py commands and exporting the model.

python -u DeepSpeech.py
–train_files /home/javi/speech/tools/backup/train/train.csv
–dev_files /home/javi/speech/tools/backup/train/dev.csv
–test_files /home/javi/speech/tools/backup/train/test.csv
–n_hidden 375
–epoch 33
–early_stop True
–earlystop_nsteps 6
–estop_mean_thresh 0.1
–estop_std_thresh 0.1
–dropout_rate 0.22
–learning_rate 0.00095
–report_count 100
–use_seq_length False
–export_dir /home/javi/speech/tools/backup/export_modal/
–checkpoint_dir /home/javi/speech/tools/backup/checkout/
–alphabet_config_path /home/javi/speech/tools/DeepSpeechWorking/data/alphabet.txt
–lm_binary_path /home/javi/speech/tools/DeepSpeechWorking/data/lm.binary
–lm_trie_path /home/javi/speech/tools/DeepSpeechWorking/data/trie
“$@”

Again, you need to be more explicit in your setup. You mention TensorFlow 1.13 and ds-ctcdecoder 0.5.1. Are you training with DeepSpeech checkout of v0.5.1 ?

Where does your language model and trie comes from ? All of that is inconsistent, and you don’t document your exact setup.

yes

I am using Kenlm and checking this link to --> TUTORIAL : How I trained a specific french model to control my robot

to create lm.binary and trie.

Could you please tell me which version is compatible for this --> deepspeech 0.6.0a4
ds-ctcdecoder 0.5.1

That tutorial is a bit old, that’s why you need to explicitely document what you do …

Could you please articulate a question here ? I’ve already explained you that those versions can’t be compatible.

1 Like

@javi.rahman So again, we do document everything on training and building a LM in the repo. Check README.md and data/lm/README.md

1 Like