Skip to content

Refactoring for use in https://github.com/mingodad/squilu #20

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
mingodad wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Refactoring for use in https://github.com/mingodad/squilu #20

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3fae82b
It seems that gcc can't optimize this so using simd here give us a go…
mingodad Jun 22, 2018
f48f825
Refactoring to be able to use it on https://github.com/mingodad/squilu
mingodad Jul 2, 2018
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
82 changes: 41 additions & 41 deletions neuralredis.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -64,7 +64,7 @@ uint64_t NRNextId = 1; /* Next neural network unique ID. */
#define NR_FLAG_BACKTRACK (1<<6) /* Auto stop with backtracking. */

/* Flags to persist when saving the NN. */
#define NR_FLAG_TO_PRESIST (NR_FLAG_REGRESSOR| \
#define NR_FLAG_TO_PERSIST (NR_FLAG_REGRESSOR| \
NR_FLAG_CLASSIFIER| \
NR_FLAG_NORMALIZE| \
NR_FLAG_OF_DETECTED)
Expand All @@ -91,7 +91,7 @@ typedef struct {
uint64_t training_max_ms; /* Max time of a single training. */
uint32_t flags; /* NR_FLAG_... */
uint32_t epochs; /* Number of training epochs so far. */
struct Ann *nn; /* Neural network structure. */
AnnRprop *nn; /* Neural network structure. */
NRDataset dataset; /* Training dataset. */
NRDataset test; /* Testing dataset. */
float dataset_error; /* Average error in the training dataset. */
Expand All @@ -104,7 +104,7 @@ typedef struct {
float *onorm; /* Outputs normalization factors. */
} NRTypeObject;

struct {
typedef struct {
RedisModuleString *key; /* Key name of the NN we are training.
Set to NULL for unused slots. */
int db_id; /* DB ID where the key is. */
Expand All @@ -115,7 +115,7 @@ struct {
float test_error; /* Test error in the last cycle. */
float class_error; /* Percentage of wrong classifications. */
int curcycle; /* Current cycle. */
} typedef NRPendingTraining;
} NRPendingTraining;

/* We take an array with NNs currently training in other threads.
* Every time an NN command is called, we try to see if there are
Expand Down Expand Up @@ -152,8 +152,8 @@ NRTypeObject *createNRTypeObject(int flags, int *layers, int numlayers, int dset
o->nn = AnnCreateNet(numlayers,layers);
o->dataset.maxlen = dset_len;
o->test.maxlen = test_len;
int ilen = INPUT_UNITS(o->nn);
int olen = OUTPUT_UNITS(o->nn);
int ilen = ANN_INPUT_UNITS(o->nn);
int olen = ANN_OUTPUT_UNITS(o->nn);
o->inorm = RedisModule_Calloc(1,sizeof(float)*ilen);
o->onorm = RedisModule_Calloc(1,sizeof(float)*olen);
for (int j = 0; j < ilen; j++) o->inorm[j] = 1;
Expand Down Expand Up @@ -216,8 +216,8 @@ void NRTypeInsertData(NRTypeObject *o, float *inputs, float *outputs,

/* Append if there is room or substitute with a random entry. */
size_t idx;
int j, numin = INPUT_UNITS(o->nn),
numout = OUTPUT_UNITS(o->nn);
int j, numin = ANN_INPUT_UNITS(o->nn),
numout = ANN_OUTPUT_UNITS(o->nn);

if (target->maxlen == target->len) {
idx = rand() % target->maxlen;
Expand Down Expand Up @@ -275,8 +275,8 @@ NRTypeObject *NRClone(NRTypeObject *o, int newid) {
copy->dataset = o->dataset;
copy->test = o->test;

int ilen = INPUT_UNITS(o->nn);
int olen = OUTPUT_UNITS(o->nn);
int ilen = ANN_INPUT_UNITS(o->nn);
int olen = ANN_OUTPUT_UNITS(o->nn);
copy->dataset.inputs = RedisModule_Alloc(sizeof(float)*ilen*o->dataset.len);
copy->dataset.outputs = RedisModule_Alloc(sizeof(float)*olen*o->dataset.len);
copy->test.inputs = RedisModule_Alloc(sizeof(float)*ilen*o->test.len);
Expand Down Expand Up @@ -318,8 +318,8 @@ void NRTransferWeights(RedisModuleCtx *ctx, NRTypeObject *dst, NRTypeObject *src
dst->test_class_error = src->test_class_error;
dst->flags |= src->flags & NR_FLAG_TO_TRANSFER;

int ilen = INPUT_UNITS(src->nn);
int olen = OUTPUT_UNITS(src->nn);
int ilen = ANN_INPUT_UNITS(src->nn);
int olen = ANN_OUTPUT_UNITS(src->nn);
memcpy(dst->inorm,src->inorm,sizeof(float)*ilen);
memcpy(dst->onorm,src->onorm,sizeof(float)*olen);
}
Expand Down Expand Up @@ -361,8 +361,8 @@ void *NRTrainingThreadMain(void *arg) {
* (NR_FLAG_CLASSIFIER), no output normalization will be done since
* the data is already in 0/1 format. */
if ((nr->flags & NR_FLAG_NORMALIZE) && nr->dataset.len) {
int ilen = INPUT_UNITS(nr->nn);
int olen = OUTPUT_UNITS(nr->nn);
int ilen = ANN_INPUT_UNITS(nr->nn);
int olen = ANN_OUTPUT_UNITS(nr->nn);
float *imax = nr->inorm;
float *omax = nr->onorm;
float *inputs = nr->dataset.inputs;
Expand Down Expand Up @@ -410,7 +410,7 @@ void *NRTrainingThreadMain(void *arg) {
}
}

struct Ann *saved = NULL; /* Saved to recover on overfitting. */
AnnRprop *saved = NULL; /* Saved to recover on overfitting. */
float saved_error; /* The test error of the saved NN. */
float saved_train_error; /* The training dataset error of the saved NN */
float saved_class_error; /* The % of classification errors of saved NN */
Expand All @@ -424,7 +424,7 @@ void *NRTrainingThreadMain(void *arg) {
0,
training_iterations,
nr->dataset.len,
NN_ALGO_BPROP);
ANN_ALGO_BPROP);
cycle_time = NRMilliseconds() - cycle_start;
nr->training_total_steps += nr->dataset.len*training_iterations;

Expand Down Expand Up @@ -747,7 +747,7 @@ int NRGenericRun_RedisCommand(RedisModuleCtx *ctx, RedisModuleString **argv, int
"Use this command with a classifier network");


int ilen = INPUT_UNITS(nr->nn);
int ilen = ANN_INPUT_UNITS(nr->nn);
if (argc != ilen+2)
return RedisModule_ReplyWithError(ctx,
"ERR number of arguments does not "
Expand All @@ -760,19 +760,19 @@ int NRGenericRun_RedisCommand(RedisModuleCtx *ctx, RedisModuleString **argv, int
"ERR invalid neural network input: must be a valid float "
"precision floating point number");
if (nr->flags & NR_FLAG_NORMALIZE) input /= nr->inorm[j];
INPUT_NODE(nr->nn,j) = input;
ANN_INPUT_NODE(nr->nn,j) = input;
}

AnnSimulate(nr->nn);

/* Output the raw net output or the class ID if the network
* is a classifier and the command invoked was NR.CLASS. */
int olen = OUTPUT_UNITS(nr->nn);
int olen = ANN_OUTPUT_UNITS(nr->nn);
if (output_class) {
float max = OUTPUT_NODE(nr->nn,0);
float max = ANN_OUTPUT_NODE(nr->nn,0);
int max_class = 0;
for(int j = 1; j < olen; j++) {
float output = OUTPUT_NODE(nr->nn,j);
float output = ANN_OUTPUT_NODE(nr->nn,j);
if (output > max) {
max = output;
max_class = j;
Expand All @@ -782,7 +782,7 @@ int NRGenericRun_RedisCommand(RedisModuleCtx *ctx, RedisModuleString **argv, int
} else {
RedisModule_ReplyWithArray(ctx,olen);
for(int j = 0; j < olen; j++) {
float output = OUTPUT_NODE(nr->nn,j);
float output = ANN_OUTPUT_NODE(nr->nn,j);
if (!(nr->flags & NR_FLAG_CLASSIFIER) &&
(nr->flags & NR_FLAG_NORMALIZE))
{
Expand Down Expand Up @@ -816,8 +816,8 @@ int NRObserve_RedisCommand(RedisModuleCtx *ctx, RedisModuleString **argv, int ar
return RedisModule_ReplyWithError(ctx,REDISMODULE_ERRORMSG_WRONGTYPE);

NRTypeObject *nr = RedisModule_ModuleTypeGetValue(key);
int ilen = INPUT_UNITS(nr->nn);
int olen = OUTPUT_UNITS(nr->nn);
int ilen = ANN_INPUT_UNITS(nr->nn);
int olen = ANN_OUTPUT_UNITS(nr->nn);
int oargs = (nr->flags & NR_FLAG_CLASSIFIER) ? 1 : olen;
int target = NR_INSERT_NO_TARGET;

Expand Down Expand Up @@ -1017,9 +1017,9 @@ int NRInfo_RedisCommand(RedisModuleCtx *ctx, RedisModuleString **argv, int argc)
RedisModule_ReplyWithLongLong(ctx,!!(nr->flags & NR_FLAG_TRAINING));

RedisModule_ReplyWithSimpleString(ctx,"layout");
RedisModule_ReplyWithArray(ctx,LAYERS(nr->nn));
for (int i = LAYERS(nr->nn)-1; i >= 0; i--) {
int units = UNITS(nr->nn,i);
RedisModule_ReplyWithArray(ctx,ANN_LAYERS(nr->nn));
for (int i = ANN_LAYERS(nr->nn)-1; i >= 0; i--) {
int units = ANN_UNITS(nr->nn,i);
if (i != 0) units--; /* Don't count the bias unit. */
RedisModule_ReplyWithLongLong(ctx,units);
}
Expand Down Expand Up @@ -1111,8 +1111,8 @@ int NRGetdata_RedisCommand(RedisModuleCtx *ctx, RedisModuleString **argv, int ar

NRTypeObject *nr = RedisModule_ModuleTypeGetValue(key);

int ilen = INPUT_UNITS(nr->nn);
int olen = OUTPUT_UNITS(nr->nn);
int ilen = ANN_INPUT_UNITS(nr->nn);
int olen = ANN_OUTPUT_UNITS(nr->nn);
NRDataset *target = NULL;
long long idx;

Expand Down Expand Up @@ -1173,15 +1173,15 @@ void NRTypeRdbSave(RedisModuleIO *rdb, void *value) {
NRTypeObject *nr = value;

/* Save the neural network layout. */
RedisModule_SaveUnsigned(rdb,LAYERS(nr->nn));
for (int j = 0; j < LAYERS(nr->nn); j++) {
int units = UNITS(nr->nn,j);
RedisModule_SaveUnsigned(rdb,ANN_LAYERS(nr->nn));
for (int j = 0; j < ANN_LAYERS(nr->nn); j++) {
int units = ANN_UNITS(nr->nn,j);
if (j != 0) units--; /* Don't count the bias unit. */
RedisModule_SaveUnsigned(rdb,units);
}

/* Save the object metadata. */
RedisModule_SaveUnsigned(rdb,nr->flags & NR_FLAG_TO_PRESIST);
RedisModule_SaveUnsigned(rdb,nr->flags & NR_FLAG_TO_PERSIST);
RedisModule_SaveUnsigned(rdb,nr->id);
RedisModule_SaveUnsigned(rdb,nr->training_total_steps);
RedisModule_SaveUnsigned(rdb,nr->training_total_ms);
Expand All @@ -1193,8 +1193,8 @@ void NRTypeRdbSave(RedisModuleIO *rdb, void *value) {

/* Save the neural network weights and biases. We start
* at layer 1 since the first layer are just outputs. */
for (int j = 1; j < LAYERS(nr->nn); j++) {
int weights = WEIGHTS(nr->nn,j);
for (int j = 1; j < ANN_LAYERS(nr->nn); j++) {
int weights = ANN_WEIGHTS(nr->nn,j);
for (int i = 0; i < weights; i++)
RedisModule_SaveFloat(rdb,nr->nn->layer[j].weight[i]);
for (int i = 0; i < weights; i++)
Expand All @@ -1204,8 +1204,8 @@ void NRTypeRdbSave(RedisModuleIO *rdb, void *value) {
}

/* Save the normalization vectors. */
uint32_t ilen = INPUT_UNITS(nr->nn);
uint32_t olen = OUTPUT_UNITS(nr->nn);
uint32_t ilen = ANN_INPUT_UNITS(nr->nn);
uint32_t olen = ANN_OUTPUT_UNITS(nr->nn);
for (uint32_t j = 0; j < ilen; j++) RedisModule_SaveFloat(rdb,nr->inorm[j]);
for (uint32_t j = 0; j < olen; j++) RedisModule_SaveFloat(rdb,nr->onorm[j]);

Expand Down Expand Up @@ -1261,8 +1261,8 @@ void *NRTypeRdbLoad(RedisModuleIO *rdb, int encver) {
nr->test_class_error = RedisModule_LoadFloat(rdb);

/* Load the neural network weights. */
for (int j = 1; j < LAYERS(nr->nn); j++) {
int weights = WEIGHTS(nr->nn,j);
for (int j = 1; j < ANN_LAYERS(nr->nn); j++) {
int weights = ANN_WEIGHTS(nr->nn,j);
for (int i = 0; i < weights; i++)
nr->nn->layer[j].weight[i] = RedisModule_LoadFloat(rdb);
for (int i = 0; i < weights; i++)
Expand All @@ -1272,8 +1272,8 @@ void *NRTypeRdbLoad(RedisModuleIO *rdb, int encver) {
}

/* Load the normalization vector. */
uint32_t ilen = INPUT_UNITS(nr->nn);
uint32_t olen = OUTPUT_UNITS(nr->nn);
uint32_t ilen = ANN_INPUT_UNITS(nr->nn);
uint32_t olen = ANN_OUTPUT_UNITS(nr->nn);
for (uint32_t j = 0; j < ilen; j++)
nr->inorm[j] = RedisModule_LoadFloat(rdb);
for (uint32_t j = 0; j < olen; j++)
Expand Down
Loading