- almost 3 years Using multiple neuron layers > 3 results in NaN
- almost 3 years Installing Torch on Virtual Machine
- almost 3 years Training and Validation not handled correctly
- almost 3 years How to install this repository as a lua package
- almost 3 years 'vocab_attr' (a nil value) Error?
- almost 3 years Is it possible to grade style similarity?
- almost 3 years question about splitTable in model/LSTM.lua
- almost 3 years how to understand the parameters of LSTM.lstm constructor
- almost 3 years Can you please share the parameters values used for Linux data in the blogpost
- about 3 years CR line terminators affecting training
- about 3 years No Luarocks module found for util.OneHot
- about 3 years Torch Issue
- about 3 years what can be some possible applications of character level language modeling ?? out of curosity
- about 3 years Question: Single line seq_length?
- over 3 years Question about the training loss and validation loss.
- over 3 years How to test the char-rnn model for some random text file? How to compile/run in ubuntu to get the next predicted character?
- over 3 years Multinomial Error
- over 3 years Endless "performing test c_has_sse4_1_2"
- over 3 years Subsequence Probability
- over 3 years Question About train.lua
- over 3 years memory problems and frequent crashes
- over 3 years Training keeps running out of memory in the "cloning rnn" stage
- over 3 years Cannot Train: attempt to call field 'ClassNLLCriterion_updateOutput' (a nil value)
- over 3 years Feature: beam search for improving global quality of new text samples
- over 3 years code for init_from in training has few bugs
- over 3 years Incorrect description of sampling temperature in Docs
- almost 4 years Add a check to see if training loss has gone to infinity
- almost 4 years Adjust learning rate based on score improvement
- almost 4 years allow the validation to be a seperate file
- almost 4 years add the ablity vary the temperature based on "scale"
- almost 4 years priming from an external file when sampling
- almost 4 years english and french text together causes crash
- almost 4 years How to use this code to model word-level RNN.
- almost 4 years Error using pre-trained model
- almost 4 years Are pre-trained models available anywhere?
- almost 4 years CPU is not being used 100%
- almost 4 years Memory Issue in Scaling sequence length
- almost 4 years bad argument #3 to 'ClassNLLCriterion_updateOutput' (torch.CudaTensor expected, got number)
- almost 4 years Out of memory error while evaluating a split
- almost 4 years Food for thought: Train data one-by-one?
- almost 4 years Problem with Torch
- almost 4 years Error: table index is nil
- almost 4 years luajit: Out of memory
- about 4 years Scoring Functionality
- about 4 years Unable to sample checkpoint trained with OpenCL
- about 4 years Non-text data in input.txt
- about 4 years Training with many short examples of variable length
- about 4 years init_from fails with any given checkpoint
- about 4 years Utf-8 support
- about 4 years Should init_from parameter start train.lua from 1 if multiple checkpoints exist?
- about 4 years train_loss and grad/param norm always are nan
- about 4 years Training w/ GPU, sampling without GPU
- about 4 years error in sample.lua: bad argument #2 to '?' (invalid multinomial distribution (sum of probabilities <= 0) at /root/torch/pkg/torch/lib/TH/generic/THTensorRandom.c:109)
- about 4 years sample.lua fails to run: error in function addmm()
- about 4 years input.txt max size
- about 4 years Training determinism
- about 4 years Word Level Encodings

- almost 3 years cunn and cutorch are not found (but they ARE installed)
- almost 3 years Can I delete older checkpoints?
- about 3 years Loss is exploding, aborting
- about 3 years usual prediciton error
- over 3 years How to use a trained char-based RNN language model?
- over 3 years Justification of cloning `rnn` and `criterion`
- over 3 years time/batch about same regardless of gpuid
- over 3 years on clone_many_times()
- over 3 years For info, started on a version of char-rnn using Element Research rnn modules
- over 3 years char-rnn is 7% slower on NVIDIA 940M since commit 82baee
- over 3 years rnn and ocr
- over 3 years learningRate not saved
- over 3 years Why the lstm is not built using the container?
- over 3 years bad argument #1 to 'set'
- over 3 years In the sampling, should we subtract the max first, for numerical stability?
- over 3 years Question: How to print out the activations of non-linear states and gates?
- over 3 years Why the seq_length and batch_size are set 50 or other value?
- over 3 years openLC error
- over 3 years Question: How to reset all states after each sequence? [Duplicate]
- over 3 years Hints for a Unified Religion Text Generator
- almost 4 years After the last commit
- almost 4 years Why do we need to carry over the rnn_state in the evaluation for batch prediction?
- almost 4 years Simple recipe to have another 6% computation speedup in CUDA mode
- almost 4 years Plausible model misuse
- almost 4 years checkpoints not saving after upgrade to gpu capability
- almost 4 years second RMSprop update yields NaN loss?
- almost 4 years Warning: comparison of integers of different signs
- almost 4 years Gradient divided by sequence length
- almost 4 years About GPU Usage and Monitoring
- almost 4 years Question: why don't we learn the initial states?
- almost 4 years How is the weighted matrix trained?
- almost 4 years Where I sholud put the download codes?
- almost 4 years newline character in prime text
- almost 4 years params_lstm:copy(x) does not work for GPU version LSTM RNN
- almost 4 years Global var 'path' is null in utils
- almost 4 years Only copy the state of top_h?
- almost 4 years Training the net locks up the operating system
- almost 4 years How do I know GPU is running instead of CPU?
- almost 4 years Peephole connections
- almost 4 years cutorch problem - invalid argument at (...)/cutorch/lib/THC/THCTensor.cu:32
- almost 4 years Question: why doesnt state get reset after each epoch?
- almost 4 years Multi-GPU support
- almost 4 years Train Error on GPU ClassNLLCriterion.lua:34: bad argument #1 (field weights does not exist)
- almost 4 years Sampling to file
- almost 4 years can we update char-rnn model ?
- almost 4 years CUDA support broken against current torch/nn
- almost 4 years Gradient magnitute depends on seq_length
- almost 4 years Error loading module 'libnn' while trying to run