Skip to content

ggml : set output of view src #25729

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

Merged
ggerganov merged 9 commits into from
Jul 28, 2026
Merged

ggml : set output of view src #25729

Show file tree
Hide file tree
Changes from 6 commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
2ee90a9
llama-graph: set_outputs to t->view_src
am17an Jul 15, 2026
fd4fd67
change set_output to GGML_ASSERT about views not being outputs
am17an Jul 16, 2026
3c430ec
Merge branch 'master' into pr/25729
ggerganov Jul 28, 2026
28a62f9
sampler : avoid views in outputs
ggerganov Jul 28, 2026
f3aa737
cont : fix dist sampler
ggerganov Jul 28, 2026
8fc77ee
cont : consistent logits handling
ggerganov Jul 28, 2026
299077b
ggml : set output of view src
gaugarg-nv Jul 28, 2026
bf17c7d
graph : simplify set_outputs()
ggerganov Jul 28, 2026
7255f0a
cont : cleanup
ggerganov Jul 28, 2026
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
25 changes: 16 additions & 9 deletions src/llama-graph.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1228,45 +1228,50 @@ void llm_graph_result::set_inputs(const llama_ubatch * ubatch) {
}

void llm_graph_result::set_outputs(const llm_graph_params & params) {
auto set_output = [](ggml_tensor * t) {
ggml_set_output(t);
GGML_ASSERT((t->view_src == nullptr || t->view_src->flags & GGML_TENSOR_FLAG_OUTPUT) && "src tensor of an output view is not an output");
};

if (t_logits != nullptr) {
ggml_set_output(t_logits);
set_output(t_logits);
}
if (t_embd != nullptr) {
ggml_set_output(t_embd);
set_output(t_embd);
}
if (t_embd_pooled != nullptr) {
ggml_set_output(t_embd_pooled);
set_output(t_embd_pooled);
}
if (t_h_nextn != nullptr) {
ggml_set_output(t_h_nextn);
set_output(t_h_nextn);
}
{
const auto & embeddings_layer_inp = params.cparams.embeddings_layer_inp;
for (size_t il = 0; il < embeddings_layer_inp.size(); ++il) {
if (embeddings_layer_inp[il]) {
GGML_ASSERT(t_layer_inp[il] != nullptr && "layer input tensor is null");
ggml_set_output(t_layer_inp[il]);
set_output(t_layer_inp[il]);
}
}
}
for (auto & [seq_id, t] : t_sampled) {
if (t != nullptr) {
ggml_set_output(t);
set_output(t);
}
}
for (auto & [seq_id, t] : t_sampled_probs) {
if (t != nullptr) {
ggml_set_output(t);
set_output(t);
}
}
for (auto & [seq_id, t] : t_sampled_logits) {
if (t != nullptr) {
ggml_set_output(t);
set_output(t);
}
}
for (auto & [seq_id, t] : t_candidates) {
if (t != nullptr) {
ggml_set_output(t);
set_output(t);
}
}
}
Expand Down Expand Up @@ -3437,7 +3442,9 @@ void llm_graph_context::build_sampling() const {
// add a dummy row of logits
// this trick makes the graph static, regardless of which samplers are activated
// this is important in order to minimize graph reallocations
// note: mark this tensor as output since there could be output views of its data
Comment thread
ggerganov marked this conversation as resolved.
Resolved
ggml_tensor * logits_t = ggml_pad(ctx0, res->t_logits, 0, 1, 0, 0);
ggml_set_output(logits_t);

for (const auto & [seq_id, sampler] : samplers) {
const auto it = seq_to_logit_row.find(seq_id);
Expand Down
57 changes: 36 additions & 21 deletions src/llama-sampler.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -993,7 +993,9 @@ static void llama_sampler_greedy_backend_apply(
GGML_UNUSED(gf);
GGML_UNUSED(smpl);

struct ggml_tensor * curl = ggml_argmax(ctx, data->logits);
struct ggml_tensor * logits = ggml_reshape_1d(ctx, data->logits, ggml_nelements(data->logits));

struct ggml_tensor * curl = ggml_argmax(ctx, logits);
ggml_set_name(curl, "greedy_argmax");

data->sampled = curl;
Expand Down Expand Up @@ -1158,7 +1160,10 @@ static void llama_sampler_dist_backend_apply(
ggml_set_name (sctx->inp_uniform, "uniform");
ggml_set_input(sctx->inp_uniform);

struct ggml_tensor * probs = ggml_soft_max(ctx, data->logits);
// flatten
struct ggml_tensor * logits = ggml_reshape_1d(ctx, data->logits, ggml_nelements(data->logits));

struct ggml_tensor * probs = ggml_soft_max(ctx, logits);
ggml_set_name(probs, "dist_probs");

struct ggml_tensor * cumsum = ggml_cumsum(ctx, probs);
Expand Down Expand Up @@ -1289,22 +1294,22 @@ static void llama_sampler_top_k_backend_apply(
struct llama_sampler_data * data) {
auto * sctx = (llama_sampler_top_k *) smpl->ctx;

struct ggml_tensor * top_k = ggml_top_k(ctx, data->logits, sctx->k);
struct ggml_tensor * logits = ggml_reshape_1d(ctx, data->logits, ggml_nelements(data->logits));

struct ggml_tensor * top_k = ggml_top_k(ctx, logits, sctx->k);
ggml_set_name(top_k, "top_k");

if (data->candidates) {
struct ggml_tensor * candidates_rows = ggml_reshape_2d(ctx, data->candidates, 1, data->candidates->ne[0]);
data->candidates = ggml_get_rows(ctx, candidates_rows, top_k);
data->candidates = ggml_reshape_1d(ctx, data->candidates, sctx->k);
ggml_set_name(data->candidates, "top_k_candidates");
} else {
data->candidates = top_k;
}

struct ggml_tensor * logits_rows = ggml_reshape_2d(ctx, data->logits, 1, data->logits->ne[0]);
struct ggml_tensor * top_k_rows = ggml_get_rows(ctx, logits_rows, top_k);
data->logits = ggml_reshape_1d(ctx, top_k_rows, sctx->k);
ggml_set_name(top_k_rows, "top_k_rows");
struct ggml_tensor * logits_rows = ggml_reshape_2d(ctx, logits, 1, logits->ne[0]);
data->logits = ggml_get_rows(ctx, logits_rows, top_k);
ggml_set_name(data->logits, "top_k_rows");

GGML_UNUSED(gf);
}
Expand Down Expand Up @@ -1435,21 +1440,25 @@ static void llama_sampler_top_p_backend_apply(
struct llama_sampler_data * data) {
auto * sctx = (llama_sampler_top_p *) smpl->ctx;

// flatten
struct ggml_tensor * logits = ggml_reshape_1d(ctx, data->logits, ggml_nelements(data->logits));

auto ggml_sort = [ctx](struct ggml_tensor * a, struct ggml_tensor * b) {
GGML_ASSERT(ggml_nrows(a) == 1);
struct ggml_tensor * a_reshaped = ggml_reshape_2d(ctx, a, 1, a->ne[0]);
struct ggml_tensor * a_sorted = ggml_get_rows(ctx, a_reshaped, b);
return ggml_reshape_1d(ctx, a_sorted, a->ne[0]);
return a_sorted;
};

// Get the sorted logits in descending order.
struct ggml_tensor * sorted_idx = ggml_argsort(ctx, data->logits, GGML_SORT_ORDER_DESC);
struct ggml_tensor * sorted_idx = ggml_argsort(ctx, logits, GGML_SORT_ORDER_DESC);
ggml_set_name(sorted_idx, "top_p_sorted_idx");

// Do the sorting via reshape + get_rows
struct ggml_tensor * sorted_logits = ggml_sort(data->logits, sorted_idx);
struct ggml_tensor * sorted_logits = ggml_sort(logits, sorted_idx);
ggml_set_name(sorted_logits, "top_p_sorted_logits");

sorted_logits = ggml_reshape_1d(ctx, sorted_logits, logits->ne[0]);
Comment thread
ggerganov marked this conversation as resolved.
Resolved
struct ggml_tensor * softmax = ggml_soft_max(ctx, sorted_logits);
ggml_set_name(softmax, "top_p_softmax");

Expand Down Expand Up @@ -1626,10 +1635,12 @@ static void llama_sampler_min_p_backend_apply(
struct llama_sampler_data * data) {
auto * sctx = (llama_sampler_min_p *) smpl->ctx;

struct ggml_tensor * max_idx = ggml_argmax(ctx, data->logits);
struct ggml_tensor * logits = ggml_reshape_1d(ctx, data->logits, ggml_nelements(data->logits));

struct ggml_tensor * max_idx = ggml_argmax(ctx, logits);
ggml_set_name(max_idx, "max_idx");

struct ggml_tensor * logits_rows = ggml_reshape_2d(ctx, data->logits, 1, data->logits->ne[0]);
struct ggml_tensor * logits_rows = ggml_reshape_2d(ctx, logits, 1, logits->ne[0]);
ggml_set_name(logits_rows, "logits_rows");

struct ggml_tensor * max_logit = ggml_get_rows(ctx, logits_rows, max_idx);
Expand All @@ -1640,7 +1651,7 @@ static void llama_sampler_min_p_backend_apply(
ggml_set_name(threshold, "min_p_threshold");

// Subtract the threshold from logits.
struct ggml_tensor * sub = ggml_sub(ctx, data->logits, threshold);
struct ggml_tensor * sub = ggml_sub(ctx, logits, threshold);

// Create a mask where logits below the threshold are 0 (discard),
// and others are 1 (keep).
Expand All @@ -1652,7 +1663,7 @@ static void llama_sampler_min_p_backend_apply(
struct ggml_tensor * min_p_bias = ggml_log(ctx, mask);
ggml_set_name(min_p_bias, "min_p_bias");

data->logits = ggml_add(ctx, data->logits, min_p_bias);
data->logits = ggml_add(ctx, logits, min_p_bias);
ggml_set_name(data->logits, "min_p_logits");

GGML_UNUSED(gf);
Expand Down Expand Up @@ -1829,18 +1840,20 @@ static void llama_sampler_backend_temp_sampling(
struct llama_sampler_data * data,
float temp) {
if (temp <= 0.0f) {
struct ggml_tensor * logits = ggml_reshape_1d(ctx, data->logits, ggml_nelements(data->logits));

// Find the most probable token index.
struct ggml_tensor * max_idx = ggml_argmax(ctx, data->logits);
struct ggml_tensor * max_idx = ggml_argmax(ctx, logits);
ggml_set_name(max_idx, "temp_max_idx");

if (data->candidates) {
struct ggml_tensor * candidates_rows = ggml_reshape_2d(ctx, data->candidates, 1, data->candidates->ne[0]);
struct ggml_tensor * candidates_rows = ggml_reshape_2d(ctx, data->candidates, 1, ggml_nelements(data->candidates));
data->candidates = ggml_get_rows(ctx, candidates_rows, max_idx);
} else {
data->candidates = max_idx;
}

struct ggml_tensor * logits_rows = ggml_reshape_2d(ctx, data->logits, 1, data->logits->ne[0]);
struct ggml_tensor * logits_rows = ggml_reshape_2d(ctx, logits, 1, ggml_nelements(logits));
data->logits = ggml_get_rows(ctx, logits_rows, max_idx);

return;
Expand Down Expand Up @@ -2019,13 +2032,15 @@ static void llama_sampler_temp_ext_backend_apply(
return;
}

struct ggml_tensor * logits = ggml_reshape_1d(ctx, data->logits, ggml_nelements(data->logits));

// Calculate min_temp, max_temp, and max_entropy.
const float min_temp = std::max(0.0f, sctx->temp - sctx->delta);
const float max_temp = sctx->temp + sctx->delta;
const float max_entropy = logf(data->logits->ne[0]);
const float max_entropy = logf(logits->ne[0]);

// Calculate the probabilities.
struct ggml_tensor * probs = ggml_soft_max(ctx, data->logits);
struct ggml_tensor * probs = ggml_soft_max(ctx, logits);
ggml_set_name(probs, "temp_ext_softmax_probs");

// Clamp probabilities to avoid log(0) which would give -inf
Expand Down Expand Up @@ -2063,7 +2078,7 @@ static void llama_sampler_temp_ext_backend_apply(
ggml_set_name(dyn_temp, "temp_ext_dyn_temp");

// Scale the logits by the dynamic temperature
struct ggml_tensor * scaled_logits = ggml_div(ctx, data->logits, dyn_temp);
struct ggml_tensor * scaled_logits = ggml_div(ctx, logits, dyn_temp);
ggml_set_name(scaled_logits, "temp_ext_scaled_logits");

data->logits = scaled_logits;
Expand Down
Loading