Skip to content

[pull] master from ggml-org:master #879

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
pull merged 1 commit into from
Feb 15, 2026
Merged

[pull] master from ggml-org:master #879

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
76 changes: 26 additions & 50 deletions convert_hf_to_gguf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2726,8 +2726,6 @@ def set_gguf_parameters(self):
super().set_gguf_parameters()

# MoE parameters
if (n_experts := self.hparams.get("num_experts")) is not None:
self.gguf_writer.add_expert_count(n_experts)
if (n_shared_experts := self.hparams.get("num_shared_experts")) is not None:
self.gguf_writer.add_expert_shared_count(n_shared_experts)
if (moe_intermediate_size := self.hparams.get("moe_intermediate_size")) is not None:
Expand All @@ -2749,7 +2747,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter
# Handle expert weights - they're already merged in the HF format
# process the experts separately
if name.find("mlp.experts") != -1:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -4197,8 +4195,6 @@ class Qwen2MoeModel(TextModel):

def set_gguf_parameters(self):
super().set_gguf_parameters()
if (n_experts := self.hparams.get("num_experts")) is not None:
self.gguf_writer.add_expert_count(n_experts)
if (moe_intermediate_size := self.hparams.get("moe_intermediate_size")) is not None:
self.gguf_writer.add_expert_feed_forward_length(moe_intermediate_size)
logger.info(f"gguf: expert feed forward length = {moe_intermediate_size}")
Expand Down Expand Up @@ -4243,7 +4239,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter
return

if name.find("experts") != -1:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -4994,13 +4990,13 @@ class PhiMoeModel(Phi3MiniModel):

def set_gguf_parameters(self):
super().set_gguf_parameters()
self.gguf_writer.add_expert_used_count(self.hparams["num_experts_per_tok"])
self.gguf_writer.add_expert_count(self.hparams["num_local_experts"])
self.gguf_writer.add_expert_used_count(self.find_hparam(["num_experts_per_tok", "num_experts_per_token"]))
self.gguf_writer.add_expert_count(self.find_hparam(["num_local_experts", "num_experts"]))

def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
# process the experts separately
if name.find("block_sparse_moe.experts") != -1:
n_experts = self.hparams["num_local_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -5412,7 +5408,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter

# process the experts separately
if name.find("block_sparse_moe.experts") != -1:
n_experts = self.find_hparam(["num_local_experts", "num_experts"], optional=False)
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -6007,12 +6003,13 @@ def modify_tensors(self, data_torch: torch.Tensor, name: str, bid: int | None) -
if "mlp.experts.bias" in name:
return # Explicitly return.

n_experts = self.find_hparam(["num_local_experts", "num_experts"])
if "mlp.experts.mlp.w1" in name:
data_torch = data_torch.view(self.hparams["num_experts"], self.hparams["n_inner"], self.hparams["n_embd"])
data_torch = data_torch.view(n_experts, self.hparams["n_inner"], self.hparams["n_embd"])
name += ".weight"

if "mlp.experts.mlp.w2" in name:
data_torch = data_torch.view(self.hparams["num_experts"], self.hparams["n_inner"], self.hparams["n_embd"])
data_torch = data_torch.view(n_experts, self.hparams["n_inner"], self.hparams["n_embd"])
data_torch = data_torch.transpose(1, 2)
name += ".weight"

Expand All @@ -6022,7 +6019,6 @@ def set_gguf_parameters(self):
super().set_gguf_parameters()
if self.is_moe:
self.gguf_writer.add_moe_every_n_layers(self.hparams["moe_every_n_layers"])
self.gguf_writer.add_expert_count(self.hparams["num_experts"])
self.gguf_writer.add_expert_used_count(self.hparams["moe_top_k"])

def _is_tokenizer_xlmroberta(self) -> bool:
Expand Down Expand Up @@ -7259,8 +7255,8 @@ def set_gguf_parameters(self):
self.gguf_writer.add_ssm_state_size(d_state)
self.gguf_writer.add_ssm_time_step_rank(dt_rank)
self.gguf_writer.add_layer_norm_rms_eps(rms_norm_eps)
self.gguf_writer.add_expert_count(self.hparams["num_experts"])
self.gguf_writer.add_expert_used_count(self.hparams["num_experts_per_tok"])
self.gguf_writer.add_expert_count(self.find_hparam(["num_local_experts", "num_experts"]))
self.gguf_writer.add_expert_used_count(self.find_hparam(["num_experts_per_tok", "num_experts_per_token"]))
self.gguf_writer.add_file_type(self.ftype)

_experts: list[dict[str, Tensor]] | None = None
Expand All @@ -7278,7 +7274,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter

# process the experts separately
if ".feed_forward.experts." in name:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])

assert bid is not None

Expand Down Expand Up @@ -7426,16 +7422,14 @@ class OlmoeModel(TextModel):
def set_gguf_parameters(self):
super().set_gguf_parameters()
self.gguf_writer.add_layer_norm_rms_eps(1e-5)
if (n_experts := self.hparams.get("num_experts")) is not None:
self.gguf_writer.add_expert_count(n_experts)

_experts: list[dict[str, Tensor]] | None = None

# Copied from: Qwen2MoeModel
def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
# process the experts separately
if name.find("experts") != -1:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -8016,10 +8010,6 @@ class MiniMaxM2Model(TextModel):
model_arch = gguf.MODEL_ARCH.MINIMAXM2
_experts_cache: dict[int, dict[str, Tensor]] = {}

def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.hparams["num_experts"] = self.hparams["num_local_experts"]

def set_gguf_parameters(self):
super().set_gguf_parameters()

Expand All @@ -8032,7 +8022,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None):

# merge expert weights
if 'experts' in name:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

expert_cache = self._experts_cache.setdefault(bid, {})
Expand Down Expand Up @@ -9237,7 +9227,6 @@ def __init__(self, *args, **kwargs):

def set_gguf_parameters(self):
super().set_gguf_parameters()
self.gguf_writer.add_expert_count(self.hparams["num_experts"])
moe_intermediate_size = self.hparams["moe_intermediate_size"]
num_shared_experts = self.hparams["num_shared_experts"]
self.gguf_writer.add_expert_feed_forward_length(moe_intermediate_size)
Expand Down Expand Up @@ -9278,7 +9267,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter
name = name.replace("e_score_correction_bias", "e_score_correction.bias")

if name.find("mlp.experts") != -1:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -9429,7 +9418,7 @@ def __init__(self, *args, **kwargs):
# case, the model architecture needs to be updated to a standard
# "granite" or "granitemoe" model
if not self._ssm_layers:
has_experts = self.find_hparam(["num_experts_per_tok"], optional=True)
has_experts = self.find_hparam(["num_experts_per_tok", "num_experts_per_token"], optional=True)
new_arch = (
gguf.MODEL_ARCH.GRANITE_MOE
if has_experts else
Expand Down Expand Up @@ -9727,7 +9716,6 @@ def set_gguf_parameters(self):
self.gguf_writer.add_vocab_size(hparams["vocab_size"])
self.gguf_writer.add_expert_feed_forward_length(hparams["moe_intermediate_size"])
self.gguf_writer.add_expert_weights_scale(1.0)
self.gguf_writer.add_expert_count(hparams["num_experts"])
self.gguf_writer.add_expert_shared_count(hparams["num_shared_experts"])
self.gguf_writer.add_expert_weights_norm(hparams["norm_topk_prob"])

Expand Down Expand Up @@ -9761,7 +9749,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter
yield from super().modify_tensors(v,self.format_tensor_name(gguf.MODEL_TENSOR.ATTN_V, bid), bid)
return
elif name.find("mlp.experts") != -1:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -9832,7 +9820,6 @@ def set_gguf_parameters(self):
self.gguf_writer.add_expert_feed_forward_length(hparams["moe_intermediate_size"])
self.gguf_writer.add_expert_shared_feed_forward_length(hparams.get("moe_shared_expert_intermediate_size", hparams["moe_intermediate_size"] * hparams["num_shared_experts"]))
self.gguf_writer.add_expert_weights_scale(hparams["routed_scaling_factor"])
self.gguf_writer.add_expert_count(hparams["num_experts"])
self.gguf_writer.add_expert_shared_count(hparams["num_shared_experts"])
self.gguf_writer.add_expert_weights_norm(hparams["norm_topk_prob"])

Expand All @@ -9843,7 +9830,7 @@ def set_gguf_parameters(self):

def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
if "mlp.experts" in name:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -9889,8 +9876,6 @@ class GroveMoeModel(TextModel):

def set_gguf_parameters(self):
super().set_gguf_parameters()
if (n_experts := self.hparams.get("num_experts")) is not None:
self.gguf_writer.add_expert_count(n_experts)
if (moe_intermediate_size := self.hparams.get("moe_intermediate_size")) is not None:
self.gguf_writer.add_expert_feed_forward_length(moe_intermediate_size)
logger.info(f"gguf: expert feed forward length = {moe_intermediate_size}")
Expand All @@ -9911,7 +9896,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter

# process the experts separately
if name.find("chunk_experts") != -1:
n_experts = self.hparams["num_experts"] // 2 # see add_experts_per_group
n_experts = self.find_hparam(["num_local_experts", "num_experts"]) // 2 # see add_experts_per_group
assert bid is not None

if self._chunk_experts is None:
Expand All @@ -9938,7 +9923,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter
else:
return
elif name.find("experts") != -1:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -10331,7 +10316,6 @@ def set_gguf_parameters(self):
super().set_gguf_parameters()
hparams = self.hparams

self.gguf_writer.add_expert_count(hparams["num_experts"])
self.gguf_writer.add_expert_shared_feed_forward_length(hparams["intermediate_size"])

moe_intermediate_size = hparams["moe_intermediate_size"]
Expand Down Expand Up @@ -10374,7 +10358,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter
return

if name.find("mlp.experts") != -1:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -10416,16 +10400,9 @@ class LLaDAMoEModel(TextModel):

def set_gguf_parameters(self):
super().set_gguf_parameters()
if (n_experts := self.hparams.get("num_experts")) is not None:
self.gguf_writer.add_expert_count(n_experts)

if (expert_intermediate_size := self.hparams.get("expert_intermediate_size")) is not None:
self.gguf_writer.add_expert_feed_forward_length(expert_intermediate_size)

# number of experts used per token (top-k)
if (n_experts_used := self.hparams.get("num_experts_per_tok")) is not None:
self.gguf_writer.add_expert_used_count(n_experts_used)

self.gguf_writer.add_mask_token_id(156895)
self.gguf_writer.add_causal_attention(False)
self.gguf_writer.add_diffusion_shift_logits(False)
Expand All @@ -10436,7 +10413,7 @@ def set_gguf_parameters(self):
def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
# process the experts separately
if name.find("experts") != -1:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down Expand Up @@ -10773,7 +10750,6 @@ def set_gguf_parameters(self):

super().set_gguf_parameters()

self.gguf_writer.add_expert_count(self.hparams["num_experts"])
self.gguf_writer.add_expert_feed_forward_length(self.hparams["moe_intermediate_size"])
self.gguf_writer.add_leading_dense_block_count(self.hparams["num_dense_layers"])
self.gguf_writer.add_expert_gating_func(gguf.ExpertGatingFuncType.SIGMOID)
Expand All @@ -10794,7 +10770,7 @@ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iter

# merge expert weights
if 'experts' in name:
n_experts = self.hparams["num_experts"]
n_experts = self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

expert_cache = self._experts_cache.setdefault(bid, {})
Expand Down Expand Up @@ -10904,9 +10880,9 @@ class SmallThinkerModel(TextModel):

def set_gguf_parameters(self):
super().set_gguf_parameters()
if (n_experts := self.hparams.get("num_experts", self.hparams.get("moe_num_primary_experts"))) is not None:
if (n_experts := self.hparams.get("moe_num_primary_experts")) is not None:
self.gguf_writer.add_expert_count(n_experts)
if (n_experts_used := self.hparams.get("num_experts_per_tok", self.hparams.get("moe_num_active_primary_experts"))) is not None:
if (n_experts_used := self.hparams.get("moe_num_active_primary_experts")) is not None:
self.gguf_writer.add_expert_used_count(n_experts_used)
if (moe_intermediate_size := self.hparams.get("moe_ffn_hidden_size")) is not None:
self.gguf_writer.add_expert_feed_forward_length(moe_intermediate_size)
Expand All @@ -10931,7 +10907,7 @@ def set_gguf_parameters(self):
def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
# process the experts separately
if name.find("experts") != -1:
n_experts = self.hparams.get("num_experts", self.hparams.get("moe_num_primary_experts"))
n_experts = self.hparams.get("moe_num_primary_experts") or self.find_hparam(["num_local_experts", "num_experts"])
assert bid is not None

if self._experts is None:
Expand Down
Loading