vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llama : more loader cleanup, better error checking

Browse source
This commit is contained in:
slaren 2024-04-03 00:46:15 +02:00
parent fe62909618
commit 31adc93486
1 changed files with 89 additions and 55 deletions
Download patch file Download diff file Expand all files Collapse all files

144
llama.cpp
View file

@ -3198,6 +3198,14 @@ struct llama_model_loader {
return weight->tensor;
}
struct ggml_tensor * require_tensor_meta(const char * name) const {
struct ggml_tensor * tensor = get_tensor_meta(name);
if (!tensor) {
throw std::runtime_error(format("%s: tensor '%s' not found", __func__, name));
}
return tensor;
}
struct ggml_tensor * get_tensor_meta(int i) const {
return get_tensor_meta(get_tensor_name(i));
}
@ -3211,7 +3219,7 @@ struct llama_model_loader {
return tensor;
}
struct ggml_tensor * create_tensor(struct ggml_context * ctx, const std::string & name, const std::vector<int64_t> & ne, bool required = true) {
const struct ggml_tensor * check_tensor_dims(const std::string & name, const std::vector<int64_t> & ne, bool required) const {
const struct ggml_tensor * cur = get_tensor_meta(name.c_str());
if (cur == NULL) {
@ -3223,8 +3231,8 @@ struct llama_model_loader {
{
bool is_ok = true;
for (size_t i = 0; i < ne.size(); ++i) {
if (ne[i] != cur->ne[i]) {
for (size_t i = 0; i < GGML_MAX_DIMS; ++i) {
if ((i < ne.size() && ne[i] != cur->ne[i]) || (i >= ne.size() && cur->ne[i] != 1)) {
is_ok = false;
break;
}
@ -3238,9 +3246,47 @@ struct llama_model_loader {
}
}
return cur;
}
struct ggml_tensor * create_tensor(struct ggml_context * ctx, const std::string & name, const std::vector<int64_t> & ne, bool required = true) {
const struct ggml_tensor * cur = check_tensor_dims(name, ne, required);
if (cur == NULL) {
return NULL;
}
return create_tensor_for(ctx, cur);
}
struct ggml_tensor * create_tensor_as_view(struct ggml_context * ctx, struct ggml_tensor * base, const std::string & name, const std::vector<int64_t> & ne, size_t offset, bool required = true) {
const struct ggml_tensor * cur = check_tensor_dims(name, ne, required);
if (cur == NULL) {
return NULL;
}
if (cur->type != base->type) {
throw std::runtime_error(format("%s: tensor '%s' has wrong type; expected %s, got %s", __func__, name.c_str(), ggml_type_name(base->type), ggml_type_name(cur->type)));
}
std::array<int64_t, GGML_MAX_DIMS> dims;
for (size_t i = 0; i < GGML_MAX_DIMS; ++i) {
dims[i] = i < ne.size() ? ne[i] : 1;
}
struct ggml_tensor * tensor = ggml_view_4d(ctx, base,
dims[0], dims[1], dims[2], dims[3],
cur->nb[1], cur->nb[2], cur->nb[3],
offset);
ggml_set_name(tensor, name.c_str());
n_created++;
return tensor;
}
void done_getting_tensors() const {
if (n_created != n_tensors) {
throw std::runtime_error(format("%s: wrong number of tensors; expected %d, got %d", __func__, n_tensors, n_created));
@ -4417,6 +4463,11 @@ static bool llm_load_tensors(
const int64_t n_vocab = hparams.n_vocab;
const int64_t n_vocab_type = hparams.n_vocab_type;
const int64_t n_ff = hparams.n_ff;
const int64_t n_expert = hparams.n_expert;
if (n_expert > 0 && hparams.n_expert_used == 0) {
throw std::runtime_error("model has expert layers but no expert layers are used");
}
GGML_ASSERT(n_embd_gqa == n_embd_k_gqa);
@ -4471,51 +4522,39 @@ static bool llm_load_tensors(
layer.ffn_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd});
layer.ffn_gate_inp = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd}, false);
if (layer.ffn_gate_inp == nullptr) {
GGML_ASSERT(hparams.n_expert == 0);
GGML_ASSERT(hparams.n_expert_used == 0);
if (n_expert == 0) {
layer.ffn_gate = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff});
layer.ffn_down = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd});
layer.ffn_up = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff});
} else {
GGML_ASSERT(hparams.n_expert > 0);
GGML_ASSERT(hparams.n_expert_used > 0);
layer.ffn_gate_inp = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert});
layer.ffn_gate_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff, hparams.n_expert}, false);
layer.ffn_gate_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff, n_expert}, false);
if (layer.ffn_gate_exps) {
layer.ffn_down_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), { n_ff, n_embd, hparams.n_expert});
layer.ffn_up_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), {n_embd, n_ff, hparams.n_expert});
layer.ffn_down_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), { n_ff, n_embd, n_expert});
layer.ffn_up_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), {n_embd, n_ff, n_expert});
} else {
// merge split expert into a single tensor
// merge split expert into a single tensor for compatibility with older models
// requires disabling mmap
ml.use_mmap = false;
ggml_type type_gate = ml.get_tensor_meta(tn(LLM_TENSOR_FFN_GATE_EXP, "weight", i, 0).c_str())->type;
ggml_type type_down = ml.get_tensor_meta(tn(LLM_TENSOR_FFN_DOWN_EXP, "weight", i, 0).c_str())->type;
ggml_type type_up = ml.get_tensor_meta(tn(LLM_TENSOR_FFN_UP_EXP, "weight", i, 0).c_str())->type;
ggml_type type_gate = ml.require_tensor_meta(tn(LLM_TENSOR_FFN_GATE_EXP, "weight", i, 0).c_str())->type;
ggml_type type_down = ml.require_tensor_meta(tn(LLM_TENSOR_FFN_DOWN_EXP, "weight", i, 0).c_str())->type;
ggml_type type_up = ml.require_tensor_meta(tn(LLM_TENSOR_FFN_UP_EXP, "weight", i, 0).c_str())->type;
layer.ffn_gate_exps = ggml_new_tensor_3d(ctx_split, type_gate, n_embd, n_ff, hparams.n_expert);
layer.ffn_down_exps = ggml_new_tensor_3d(ctx_split, type_down, n_ff, n_embd, hparams.n_expert);
layer.ffn_up_exps = ggml_new_tensor_3d(ctx_split, type_up, n_embd, n_ff, hparams.n_expert);
layer.ffn_gate_exps = ggml_new_tensor_3d(ctx_split, type_gate, n_embd, n_ff, n_expert);
layer.ffn_down_exps = ggml_new_tensor_3d(ctx_split, type_down, n_ff, n_embd, n_expert);
layer.ffn_up_exps = ggml_new_tensor_3d(ctx_split, type_up, n_embd, n_ff, n_expert);
ggml_set_name(layer.ffn_gate_exps, tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i).c_str());
ggml_set_name(layer.ffn_down_exps, tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i).c_str());
ggml_set_name(layer.ffn_up_exps, tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i).c_str());
for (uint32_t x = 0; x < hparams.n_expert; ++x) {
for (uint32_t x = 0; x < n_expert; ++x) {
// the individual experts are loaded into a view of the merged tensor
ggml_tensor * ffn_gate_exp = ggml_view_2d(ctx_split, layer.ffn_gate_exps, n_embd, n_ff, layer.ffn_gate_exps->nb[1], layer.ffn_gate_exps->nb[2]*x);
ggml_tensor * ffn_down_exp = ggml_view_2d(ctx_split, layer.ffn_down_exps, n_ff, n_embd, layer.ffn_down_exps->nb[1], layer.ffn_down_exps->nb[2]*x);
ggml_tensor * ffn_up_exp = ggml_view_2d(ctx_split, layer.ffn_up_exps, n_embd, n_ff, layer.ffn_up_exps->nb[1], layer.ffn_up_exps->nb[2]*x);
ggml_set_name(ffn_gate_exp, tn(LLM_TENSOR_FFN_GATE_EXP, "weight", i, x).c_str());
ggml_set_name(ffn_down_exp, tn(LLM_TENSOR_FFN_DOWN_EXP, "weight", i, x).c_str());
ggml_set_name(ffn_up_exp, tn(LLM_TENSOR_FFN_UP_EXP, "weight", i, x).c_str());
ml.n_created += 3;
ml.create_tensor_as_view(ctx_split, layer.ffn_gate_exps, tn(LLM_TENSOR_FFN_GATE_EXP, "weight", i, x), { n_embd, n_ff }, layer.ffn_gate_exps->nb[2]*x);
ml.create_tensor_as_view(ctx_split, layer.ffn_down_exps, tn(LLM_TENSOR_FFN_DOWN_EXP, "weight", i, x), { n_ff, n_embd }, layer.ffn_down_exps->nb[2]*x);
ml.create_tensor_as_view(ctx_split, layer.ffn_up_exps, tn(LLM_TENSOR_FFN_UP_EXP, "weight", i, x), { n_embd, n_ff }, layer.ffn_up_exps->nb[2]*x);
}
}
}
@ -4523,6 +4562,10 @@ static bool llm_load_tensors(
} break;
case LLM_ARCH_GROK:
{
if (n_expert == 0) {
throw std::runtime_error("Grok model cannot have zero experts");
}
model.tok_embd = ml.create_tensor(ctx_input, tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab});
// output
@ -4554,43 +4597,34 @@ static bool llm_load_tensors(
layer.ffn_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd});
layer.ffn_gate_inp = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd});
layer.ffn_gate_inp = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_FFN_GATE_INP, "weight", i), {n_embd, n_expert});
GGML_ASSERT(hparams.n_expert > 0);
GGML_ASSERT(hparams.n_expert_used > 0);
layer.ffn_gate_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff, hparams.n_expert}, false);
layer.ffn_gate_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i), {n_embd, n_ff, n_expert}, false);
if (layer.ffn_gate_exps) {
layer.ffn_down_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), { n_ff, n_embd, hparams.n_expert});
layer.ffn_up_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), {n_embd, n_ff, hparams.n_expert});
layer.ffn_down_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i), { n_ff, n_embd, n_expert});
layer.ffn_up_exps = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i), {n_embd, n_ff, n_expert});
} else {
// merge split expert into a single tensor
// merge split expert into a single tensor for compatibility with older models
// requires disabling mmap
ml.use_mmap = false;
ggml_type type_gate = ml.get_tensor_meta(tn(LLM_TENSOR_FFN_GATE_EXP, "weight", i, 0).c_str())->type;
ggml_type type_down = ml.get_tensor_meta(tn(LLM_TENSOR_FFN_DOWN_EXP, "weight", i, 0).c_str())->type;
ggml_type type_up = ml.get_tensor_meta(tn(LLM_TENSOR_FFN_UP_EXP, "weight", i, 0).c_str())->type;
ggml_type type_gate = ml.require_tensor_meta(tn(LLM_TENSOR_FFN_GATE_EXP, "weight", i, 0).c_str())->type;
ggml_type type_down = ml.require_tensor_meta(tn(LLM_TENSOR_FFN_DOWN_EXP, "weight", i, 0).c_str())->type;
ggml_type type_up = ml.require_tensor_meta(tn(LLM_TENSOR_FFN_UP_EXP, "weight", i, 0).c_str())->type;
layer.ffn_gate_exps = ggml_new_tensor_3d(ctx_split, type_gate, n_embd, n_ff, hparams.n_expert);
layer.ffn_down_exps = ggml_new_tensor_3d(ctx_split, type_down, n_ff, n_embd, hparams.n_expert);
layer.ffn_up_exps = ggml_new_tensor_3d(ctx_split, type_up, n_embd, n_ff, hparams.n_expert);
layer.ffn_gate_exps = ggml_new_tensor_3d(ctx_split, type_gate, n_embd, n_ff, n_expert);
layer.ffn_down_exps = ggml_new_tensor_3d(ctx_split, type_down, n_ff, n_embd, n_expert);
layer.ffn_up_exps = ggml_new_tensor_3d(ctx_split, type_up, n_embd, n_ff, n_expert);
ggml_set_name(layer.ffn_gate_exps, tn(LLM_TENSOR_FFN_GATE_EXPS, "weight", i).c_str());
ggml_set_name(layer.ffn_down_exps, tn(LLM_TENSOR_FFN_DOWN_EXPS, "weight", i).c_str());
ggml_set_name(layer.ffn_up_exps, tn(LLM_TENSOR_FFN_UP_EXPS, "weight", i).c_str());
for (uint32_t x = 0; x < hparams.n_expert; ++x) {
for (uint32_t x = 0; x < n_expert; ++x) {
// the individual experts are loaded into a view of the merged tensor
ggml_tensor * ffn_gate_exp = ggml_view_2d(ctx_split, layer.ffn_gate_exps, n_embd, n_ff, layer.ffn_gate_exps->nb[1], layer.ffn_gate_exps->nb[2]*x);
ggml_tensor * ffn_down_exp = ggml_view_2d(ctx_split, layer.ffn_down_exps, n_ff, n_embd, layer.ffn_down_exps->nb[1], layer.ffn_down_exps->nb[2]*x);
ggml_tensor * ffn_up_exp = ggml_view_2d(ctx_split, layer.ffn_up_exps, n_embd, n_ff, layer.ffn_up_exps->nb[1], layer.ffn_up_exps->nb[2]*x);
ggml_set_name(ffn_gate_exp, tn(LLM_TENSOR_FFN_GATE_EXP, "weight", i, x).c_str());
ggml_set_name(ffn_down_exp, tn(LLM_TENSOR_FFN_DOWN_EXP, "weight", i, x).c_str());
ggml_set_name(ffn_up_exp, tn(LLM_TENSOR_FFN_UP_EXP, "weight", i, x).c_str());
ml.n_created += 3;
ml.create_tensor_as_view(ctx_split, layer.ffn_gate_exps, tn(LLM_TENSOR_FFN_GATE_EXP, "weight", i, x), { n_embd, n_ff }, layer.ffn_gate_exps->nb[2]*x);
ml.create_tensor_as_view(ctx_split, layer.ffn_down_exps, tn(LLM_TENSOR_FFN_DOWN_EXP, "weight", i, x), { n_ff, n_embd }, layer.ffn_down_exps->nb[2]*x);
ml.create_tensor_as_view(ctx_split, layer.ffn_up_exps, tn(LLM_TENSOR_FFN_UP_EXP, "weight", i, x), { n_embd, n_ff }, layer.ffn_up_exps->nb[2]*x);
}
}