Commit 48de2a1bc for llama.cpp
commit 48de2a1bcb5b8fdbc8dc41b4c562850ab8d528d2
Author: bosh <98094229+boshjerns@users.noreply.github.com>
Date: Wed Sep 30 03:33:05 2026 +0700
model : support classifier_pooling for rerankers (#29627)
* model : support classifier_pooling for ModernBERT rerankers
Assisted-by: Claude Opus 5.5
* model : read classifier pooling type in load_hparams
Write classifier.pooling_type from _try_set_pooling_type whenever the
config has classifier_pooling, and read it in
llama_model_base::load_hparams. ModernBERT falls back to mean when it
is unspecified.
Assisted-by: Claude Opus 5.5
* conversion : only accept cls and mean for classifier_pooling
Assisted-by: Claude Opus 5.5
* model : rename classifier_pooling_type to pooling_type_cls
Assisted-by: Claude Opus 5.5
diff --git a/conversion/base.py b/conversion/base.py
index 5561481e7..221aa8093 100644
--- a/conversion/base.py
+++ b/conversion/base.py
@@ -2326,6 +2326,12 @@ class TextModel(ModelBase):
raise NotImplementedError("Only MEAN, CLS, and LAST pooling types supported")
self.gguf_writer.add_pooling_type(pooling_type)
+ # pooling before a classification head (e.g. ModernBertForSequenceClassification)
+ if (classifier_pooling := self.hparams.get("classifier_pooling")) is not None:
+ if classifier_pooling not in ("cls", "mean"):
+ raise NotImplementedError(f"Unsupported classifier_pooling: {classifier_pooling}")
+ self.gguf_writer.add_classifier_pooling_type(mode_mapping[classifier_pooling])
+
def _set_vocab_glmedge(self):
from transformers import AutoTokenizer
tokenizer = AutoTokenizer.from_pretrained(self.dir_model)
diff --git a/gguf-py/gguf/constants.py b/gguf-py/gguf/constants.py
index 9a7a5e5bf..e193e9995 100644
--- a/gguf-py/gguf/constants.py
+++ b/gguf-py/gguf/constants.py
@@ -313,6 +313,7 @@ class Keys:
class Classifier:
OUTPUT_LABELS = "{arch}.classifier.output_labels"
+ POOLING_TYPE = "{arch}.classifier.pooling_type"
class ShortConv:
L_CACHE = "{arch}.shortconv.l_cache"
diff --git a/gguf-py/gguf/gguf_writer.py b/gguf-py/gguf/gguf_writer.py
index cf7b367e5..543d9b744 100644
--- a/gguf-py/gguf/gguf_writer.py
+++ b/gguf-py/gguf/gguf_writer.py
@@ -1331,6 +1331,9 @@ class GGUFWriter:
def add_classifier_output_labels(self, labels: Sequence[str]) -> None:
self.add_array(Keys.Classifier.OUTPUT_LABELS.format(arch=self.arch), labels)
+ def add_classifier_pooling_type(self, value: PoolingType) -> None:
+ self.add_uint32(Keys.Classifier.POOLING_TYPE.format(arch=self.arch), value.value)
+
# for vision models
def add_clip_has_vision_encoder(self, value: bool) -> None:
diff --git a/src/llama-arch.cpp b/src/llama-arch.cpp
index 8f1e239da..2c5a228f3 100644
--- a/src/llama-arch.cpp
+++ b/src/llama-arch.cpp
@@ -361,6 +361,7 @@ static const std::map<llm_kv, const char *> LLM_KV_NAMES = {
{ LLM_KV_CONVNEXT_BLOCK_COUNT, "%s.convnext.block_count" },
{ LLM_KV_CLASSIFIER_OUTPUT_LABELS, "%s.classifier.output_labels" },
+ { LLM_KV_CLASSIFIER_POOLING_TYPE, "%s.classifier.pooling_type" },
{ LLM_KV_TARGET_LAYERS, "%s.target_layers" },
{ LLM_KV_TARGET_HIDDEN_SIZE, "%s.target_hidden_size" },
diff --git a/src/llama-arch.h b/src/llama-arch.h
index 23b6b3810..3b8bd6428 100644
--- a/src/llama-arch.h
+++ b/src/llama-arch.h
@@ -407,6 +407,7 @@ enum llm_kv {
LLM_KV_CONVNEXT_BLOCK_COUNT,
LLM_KV_CLASSIFIER_OUTPUT_LABELS,
+ LLM_KV_CLASSIFIER_POOLING_TYPE,
LLM_KV_TARGET_LAYERS,
LLM_KV_TARGET_HIDDEN_SIZE,
diff --git a/src/llama-graph.cpp b/src/llama-graph.cpp
index a806126ef..abf3069bb 100644
--- a/src/llama-graph.cpp
+++ b/src/llama-graph.cpp
@@ -3727,8 +3727,8 @@ void llm_graph_context::build_pooling(
} break;
case LLAMA_POOLING_TYPE_RANK:
{
- if (arch == LLM_ARCH_MODERN_BERT) {
- // modern bert gte reranker builds mean first then applies prediction head and classifier
+ if (hparams.pooling_type_cls == LLAMA_POOLING_TYPE_MEAN) {
+ // modern bert with classifier_pooling = "mean" builds mean first then applies prediction head and classifier
// https://github.com/huggingface/transformers/blob/main/src/transformers/models/modernbert/modular_modernbert.py#L1404-1411
ggml_tensor * inp_mean = build_inp_mean();
cur = ggml_mul_mat(ctx0, ggml_cont(ctx0, ggml_transpose(ctx0, inp)), inp_mean);
diff --git a/src/llama-hparams.h b/src/llama-hparams.h
index 73dffcc9f..d8fbfbdd8 100644
--- a/src/llama-hparams.h
+++ b/src/llama-hparams.h
@@ -350,6 +350,7 @@ struct llama_hparams {
uint32_t dec_n_layer = 0;
enum llama_pooling_type pooling_type = LLAMA_POOLING_TYPE_NONE;
+ enum llama_pooling_type pooling_type_cls = LLAMA_POOLING_TYPE_UNSPECIFIED; // pooling before the classifier head (RANK)
enum llama_rope_type rope_type = LLAMA_ROPE_TYPE_NONE;
enum llama_rope_scaling_type rope_scaling_type_train = LLAMA_ROPE_SCALING_TYPE_NONE;
diff --git a/src/llama-model.cpp b/src/llama-model.cpp
index ab5e744b5..151a3a2a8 100644
--- a/src/llama-model.cpp
+++ b/src/llama-model.cpp
@@ -1320,6 +1320,7 @@ void llama_model_base::load_hparams(llama_model_loader & ml) {
ml.get_key(LLM_KV_EMBEDDING_LENGTH_OUT, hparams.n_embd_out_impl, false);
ml.get_key(LLM_KV_ATTENTION_CAUSAL, hparams.causal_attn, false);
ml.get_key(LLM_KV_POOLING_TYPE, hparams.pooling_type, false);
+ ml.get_key(LLM_KV_CLASSIFIER_POOLING_TYPE, hparams.pooling_type_cls, false);
ml.get_key(LLM_KV_BLOCK_COUNT, hparams.n_layer_all);
GGML_ASSERT(hparams.n_layer_all > 0 && hparams.n_layer_all <= LLAMA_MAX_LAYERS);
ml.get_key(LLM_KV_NEXTN_PREDICT_LAYERS, hparams.n_layer_nextn, false);
diff --git a/src/models/modern-bert.cpp b/src/models/modern-bert.cpp
index b7542d59b..158e3160c 100644
--- a/src/models/modern-bert.cpp
+++ b/src/models/modern-bert.cpp
@@ -20,6 +20,11 @@ void llama_model_modern_bert::load_arch_hparams(llama_model_loader & ml) {
hparams.llm_ffn_op = llm_ffn_op_type_from_string(hidden_act, LLM_FFN_GEGLU);
}
+ // GGUFs without a classifier pooling type use mean (gte-reranker-modernbert-base)
+ if (hparams.pooling_type_cls == LLAMA_POOLING_TYPE_UNSPECIFIED) {
+ hparams.pooling_type_cls = LLAMA_POOLING_TYPE_MEAN;
+ }
+
switch (hparams.n_layer()) {
case 12:
type = LLM_TYPE_47M; break; // granite-embedding-small