Additional KL-divergence statistics (#5081)

* perplexity: add top-token probability

* perplexity: add additional KL-divergence statistics

* perplexity: a better organized KL-divergence statistics output

---------

Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow 2024-01-23 15:17:20 +02:00 committed by GitHub
parent 9ecdd12e95
commit 44879ee885
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -222,13 +222,18 @@ struct kl_divergence_result {
double sum_kld2 = 0; double sum_kld2 = 0;
double sum_nll_diff = 0; double sum_nll_diff = 0;
double sum_nll_diff2 = 0; double sum_nll_diff2 = 0;
size_t n_same_top = 0;
size_t count = 0; size_t count = 0;
}; };
static void log_softmax(int n_vocab, const float * logits, const uint16_t * base_log_prob, int tok, kl_divergence_result & kld) { static double log_softmax(int n_vocab, const float * logits, const uint16_t * base_log_prob, int tok, kl_divergence_result & kld) {
float max_logit = logits[0]; float max_logit = logits[0];
int imax = 0;
for (int i = 1; i < n_vocab; ++i) { for (int i = 1; i < n_vocab; ++i) {
max_logit = std::max(max_logit, logits[i]); if (logits[i] > max_logit) {
max_logit = logits[i];
imax = i;
}
} }
double sum_exp = 0.0; double sum_exp = 0.0;
for (int i = 0; i < n_vocab; ++i) { for (int i = 0; i < n_vocab; ++i) {
@ -247,8 +252,14 @@ static void log_softmax(int n_vocab, const float * logits, const uint16_t * base
kld.sum_nll_diff2 += nll*nll; kld.sum_nll_diff2 += nll*nll;
max_logit += log_sum_exp; max_logit += log_sum_exp;
double sum = 0; double sum = 0;
int imax_base = -1;
float p_log_base_max = 0;
for (int i = 0; i < n_vocab; ++i) { for (int i = 0; i < n_vocab; ++i) {
const float p_log_base = scale*base_log_prob[i] + min_log_prob; const float p_log_base = scale*base_log_prob[i] + min_log_prob;
if (i == 0 || p_log_base > p_log_base_max) {
p_log_base_max = p_log_base;
imax_base = i;
}
if (p_log_base > -16.f) { if (p_log_base > -16.f) {
const float p_base = expf(p_log_base); const float p_base = expf(p_log_base);
sum += p_base * (p_log_base - logits[i] + max_logit); sum += p_base * (p_log_base - logits[i] + max_logit);
@ -257,14 +268,17 @@ static void log_softmax(int n_vocab, const float * logits, const uint16_t * base
kld.sum_kld += sum; kld.sum_kld += sum;
kld.sum_kld2 += sum*sum; kld.sum_kld2 += sum*sum;
++kld.count; ++kld.count;
if (imax == imax_base) ++kld.n_same_top;
return sum;
} }
static void process_logits(int n_vocab, const float * logits, const int * tokens, int n_token, static void process_logits(int n_vocab, const float * logits, const int * tokens, int n_token,
std::vector<std::thread> & workers, const std::vector<uint16_t> & base_log_probs, kl_divergence_result & kld) { std::vector<std::thread> & workers, const std::vector<uint16_t> & base_log_probs, kl_divergence_result & kld,
float * kld_values) {
std::mutex mutex; std::mutex mutex;
const int nv = 2*((n_vocab + 1)/2) + 4; const int nv = 2*((n_vocab + 1)/2) + 4;
int counter = 0; int counter = 0;
auto compute = [&mutex, &counter, &base_log_probs, &kld, n_vocab, logits, tokens, n_token, nv] () { auto compute = [&mutex, &counter, &base_log_probs, &kld, n_vocab, logits, tokens, n_token, nv, kld_values] () {
kl_divergence_result local_kld; kl_divergence_result local_kld;
while (true) { while (true) {
std::unique_lock<std::mutex> lock(mutex); std::unique_lock<std::mutex> lock(mutex);
@ -276,11 +290,13 @@ static void process_logits(int n_vocab, const float * logits, const int * tokens
kld.sum_kld2 += local_kld.sum_kld2; kld.sum_kld2 += local_kld.sum_kld2;
kld.sum_nll_diff += local_kld.sum_nll_diff; kld.sum_nll_diff += local_kld.sum_nll_diff;
kld.sum_nll_diff2 += local_kld.sum_nll_diff2; kld.sum_nll_diff2 += local_kld.sum_nll_diff2;
kld.n_same_top += local_kld.n_same_top;
kld.count += local_kld.count; kld.count += local_kld.count;
break; break;
} }
lock.unlock(); lock.unlock();
log_softmax(n_vocab, logits + i*n_vocab, base_log_probs.data() + i*nv, tokens[i+1], local_kld); double v = log_softmax(n_vocab, logits + i*n_vocab, base_log_probs.data() + i*nv, tokens[i+1], local_kld);
kld_values[i] = (float)v;
} }
}; };
for (auto & w : workers) { for (auto & w : workers) {
@ -1615,7 +1631,7 @@ static void kl_divergence(llama_context * ctx, const gpt_params & params) {
in.read((char *)&n_vocab, sizeof(n_vocab)); in.read((char *)&n_vocab, sizeof(n_vocab));
in.read((char *)&n_chunk, sizeof(n_chunk)); in.read((char *)&n_chunk, sizeof(n_chunk));
if (in.fail()) { if (in.fail()) {
fprintf(stderr, "%s: failed rwading n_vocab, n_chunk from %s\n", __func__, params.logits_file.c_str()); fprintf(stderr, "%s: failed reading n_vocab, n_chunk from %s\n", __func__, params.logits_file.c_str());
return; return;
} }
if (n_vocab != llama_n_vocab(llama_get_model(ctx))) { if (n_vocab != llama_n_vocab(llama_get_model(ctx))) {
@ -1634,6 +1650,7 @@ static void kl_divergence(llama_context * ctx, const gpt_params & params) {
const bool add_bos = llama_should_add_bos_token(llama_get_model(ctx)); const bool add_bos = llama_should_add_bos_token(llama_get_model(ctx));
std::vector<uint16_t> log_probs_uint16(size_t(n_ctx - 1 - n_ctx/2) * nv); std::vector<uint16_t> log_probs_uint16(size_t(n_ctx - 1 - n_ctx/2) * nv);
std::vector<float> kld_values(size_t(n_ctx - 1 - n_ctx/2)*n_chunk);
std::vector<float> logits; std::vector<float> logits;
if (num_batches > 1) { if (num_batches > 1) {
logits.reserve(n_ctx * n_vocab); logits.reserve(n_ctx * n_vocab);
@ -1652,6 +1669,7 @@ static void kl_divergence(llama_context * ctx, const gpt_params & params) {
}; };
kl_divergence_result kld; kl_divergence_result kld;
auto kld_ptr = kld_values.data();
for (int i = 0; i < n_chunk; ++i) { for (int i = 0; i < n_chunk; ++i) {
const int start = i * n_ctx; const int start = i * n_ctx;
@ -1705,20 +1723,24 @@ static void kl_divergence(llama_context * ctx, const gpt_params & params) {
} }
fprintf(stderr, "%.2f minutes\n", total_seconds / 60.0); fprintf(stderr, "%.2f minutes\n", total_seconds / 60.0);
printf("\nchunk PPL ln(PPL(Q)/PPL(base)) KL-Divergence\n"); printf("\nchunk PPL ln(PPL(Q)/PPL(base)) KL-Divergence Same top\n");
} }
const int first = n_ctx/2; const int first = n_ctx/2;
const float * all_logits = num_batches > 1 ? logits.data() : llama_get_logits(ctx); const float * all_logits = num_batches > 1 ? logits.data() : llama_get_logits(ctx);
process_logits(n_vocab, all_logits + first*n_vocab, tokens.data() + start + first, n_ctx - 1 - first, process_logits(n_vocab, all_logits + first*n_vocab, tokens.data() + start + first, n_ctx - 1 - first,
workers, log_probs_uint16, kld); workers, log_probs_uint16, kld, kld_ptr);
kld_ptr += n_ctx - 1 - first;
auto ppl = mean_and_uncertainty(kld.sum_nll, kld.sum_nll2, kld.count); auto ppl = mean_and_uncertainty(kld.sum_nll, kld.sum_nll2, kld.count);
auto log_ppl_ratio = mean_and_uncertainty(kld.sum_nll_diff, kld.sum_nll_diff2, kld.count); auto log_ppl_ratio = mean_and_uncertainty(kld.sum_nll_diff, kld.sum_nll_diff2, kld.count);
auto kl_div = mean_and_uncertainty(kld.sum_kld, kld.sum_kld2, kld.count); auto kl_div = mean_and_uncertainty(kld.sum_kld, kld.sum_kld2, kld.count);
auto p_top = 1.*kld.n_same_top/kld.count;
auto d_p_top = sqrt(p_top*(1 - p_top)/(kld.count - 1));
printf("%4d %10.4lf %10.5lf ± %10.5f %10.5f ± %10.5lf\n", i+1, exp(ppl.first), printf("%4d %10.4lf %10.5lf ± %10.5f %10.5f ± %10.5lf %.5f ± %.5f\n", i+1, exp(ppl.first),
log_ppl_ratio.first, log_ppl_ratio.second, kl_div.first, kl_div.second); log_ppl_ratio.first, log_ppl_ratio.second, kl_div.first, kl_div.second,
p_top, d_p_top);
fflush(stdout); fflush(stdout);
@ -1726,6 +1748,35 @@ static void kl_divergence(llama_context * ctx, const gpt_params & params) {
} }
printf("\n"); printf("\n");
if (kld.count < 100) return; // we do not wish to do statistics on so few values
std::sort(kld_values.begin(), kld_values.end());
printf("===== KL-divergence statistics\n");
auto kl_div = mean_and_uncertainty(kld.sum_kld, kld.sum_kld2, kld.count);
printf("Average: %10.6f ±%10.6lf\n", kl_div.first, kl_div.second);
auto kld_median = kld_values.size()%2 == 0 ? 0.5f*(kld_values[kld_values.size()/2] + kld_values[kld_values.size()/2-1])
: kld_values[kld_values.size()/2];
printf("Median : %10.6f\n", kld_median);
auto percentile = [&kld_values] (float fraction) {
if (fraction <= 0) return kld_values.front();
if (fraction >= 1) return kld_values.back();
float p = fraction*(kld_values.size() - 1);
size_t ip = size_t(p); p -= ip;
return (1 - p)*kld_values[ip] + p*kld_values[std::min(ip+1, kld_values.size()-1)];
};
printf("Maximum: %10.6f\n", kld_values.back());
printf("KLD_99 : %10.6f\n", percentile(0.99f));
printf("KLD_95 : %10.6f\n", percentile(0.95f));
printf("KLD_90 : %10.6f\n", percentile(0.90f));
printf("Minimum: %10.6f\n", kld_values.front());
printf("KLD_01 : %10.6f\n", percentile(0.01f));
printf("KLD_05 : %10.6f\n", percentile(0.05f));
printf("KLD_10 : %10.6f\n", percentile(0.10f));
} }
int main(int argc, char ** argv) { int main(int argc, char ** argv) {