yield from super().modify_tensors(data_torch, name, bid)
+@ModelBase.register("Ernie4_5_ForCausalLM")
+class Ernie4_5Model(TextModel):
+ model_arch = gguf.MODEL_ARCH.ERNIE4_5
+
+ def set_vocab(self):
+ self._set_vocab_sentencepiece()
+
+ def set_gguf_parameters(self):
+ super().set_gguf_parameters()
+
+ def modify_tensors(self, data_torch: Tensor, name: str, bid: int | None) -> Iterable[tuple[str, Tensor]]:
+ num_heads = self.hparams["num_attention_heads"]
+ num_kv_heads = self.hparams["num_key_value_heads"]
+ head_dim = self.hparams["head_dim"]
+
+ if "ernie." in name:
+ name = name.replace("ernie.", "model.")
+ # split the qkv weights
+ # qkv_proj shape: [(num_heads + 2 * num_kv_heads) * head_dim, hidden_size]
+ if "qkv_proj" in name:
+ name_q = name.replace("qkv_proj.weight", "q_proj.weight")
+ name_k = name.replace("qkv_proj.weight", "k_proj.weight")
+ name_v = name.replace("qkv_proj.weight", "v_proj.weight")
+ total_q_dim = num_heads * head_dim
+ total_k_dim = num_kv_heads * head_dim
+ total_v_dim = num_kv_heads * head_dim
+ q_proj_weight, k_proj_weight, v_proj_weight = data_torch.split([total_q_dim, total_k_dim, total_v_dim], dim=0)
+ return [
+ (self.map_tensor_name(name_q), q_proj_weight),
+ (self.map_tensor_name(name_k), k_proj_weight),
+ (self.map_tensor_name(name_v), v_proj_weight)
+ ]
+ # split the up_gate_proj into gate and up
+ # up_gate_proj shape: [2 * intermediate_size, hidden_size]
+ if "up_gate_proj" in name:
+ name_up = name.replace("up_gate_proj.weight", "up_proj.weight")
+ name_gate = name.replace("up_gate_proj.weight", "gate_proj.weight")
+ dim_half = data_torch.shape[0] // 2
+ gate_proj_weight, up_proj_weight = data_torch.split(dim_half, dim=0)
+ return [
+ (self.map_tensor_name(name_gate), gate_proj_weight),
+ (self.map_tensor_name(name_up), up_proj_weight)
+ ]
+ return [(self.map_tensor_name(name), data_torch)]
+
+
@ModelBase.register(
"Qwen2VLModel",
"Qwen2VLForConditionalGeneration",
case LLM_TYPE_475M: return "475M";
case LLM_TYPE_770M: return "770M";
case LLM_TYPE_780M: return "780M";
+ case LLM_TYPE_0_3B: return "0.3B";
case LLM_TYPE_0_5B: return "0.5B";
case LLM_TYPE_0_6B: return "0.6B";
case LLM_TYPE_1B: return "1B";
default: type = LLM_TYPE_UNKNOWN;
}
} break;
+ case LLM_ARCH_ERNIE4_5:
+ {
+ ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
+ switch (hparams.n_layer) {
+ case 18: type = LLM_TYPE_0_3B; break;
+ default: type = LLM_TYPE_UNKNOWN;
+ }
+ } break;
default: throw std::runtime_error("unsupported model architecture");
}
layer.rope_freqs = create_tensor(tn(LLM_TENSOR_ROPE_FREQS, "weight", i), {n_rot/2}, TENSOR_NOT_REQUIRED | (i != 0 ? TENSOR_DUPLICATED : 0));
+ layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0);
+ layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0);
+ }
+ } break;
+ case LLM_ARCH_ERNIE4_5:
+ {
+ tok_embd = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, 0);
+
+ // output
+ output_norm = create_tensor(tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd}, 0);
+ output = create_tensor(tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, TENSOR_NOT_REQUIRED);
+ // if output is NULL, init from the input tok embed
+ if (output == NULL) {
+ output = create_tensor(tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, TENSOR_DUPLICATED);
+ }
+
+ for (int i = 0; i < n_layer; ++i) {
+ auto & layer = layers[i];
+
+ layer.attn_norm = create_tensor(tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd}, 0);
+
+ layer.wq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd_head_k * n_head}, 0);
+ layer.wk = create_tensor(tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa}, 0);
+ layer.wv = create_tensor(tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa}, 0);
+ layer.wo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd_head_k * n_head, n_embd}, 0);
+
+ // optional bias tensors
+ layer.bq = create_tensor(tn(LLM_TENSOR_ATTN_Q, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED);
+ layer.bk = create_tensor(tn(LLM_TENSOR_ATTN_K, "bias", i), {n_embd_gqa}, TENSOR_NOT_REQUIRED);
+ layer.bv = create_tensor(tn(LLM_TENSOR_ATTN_V, "bias", i), {n_embd_gqa}, TENSOR_NOT_REQUIRED);
+ layer.bo = create_tensor(tn(LLM_TENSOR_ATTN_OUT, "bias", i), {n_embd}, TENSOR_NOT_REQUIRED);
+
+ layer.ffn_norm = create_tensor(tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd}, 0);
+ layer.ffn_gate = create_tensor(tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff}, 0);
layer.ffn_down = create_tensor(tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd}, 0);
layer.ffn_up = create_tensor(tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff}, 0);
}
}
};
+struct llm_build_ernie4_5 : public llm_graph_context {
+ llm_build_ernie4_5(const llama_model & model, const llm_graph_params & params, ggml_cgraph * gf) : llm_graph_context(params) {
+ const int64_t n_embd_head = hparams.n_embd_head_v;
+
+ GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
+ GGML_ASSERT(n_embd_head == hparams.n_rot);
+
+ ggml_tensor * cur;
+ ggml_tensor * inpL;
+
+ inpL = build_inp_embd(model.tok_embd);
+
+ // inp_pos - contains the positions
+ ggml_tensor * inp_pos = build_inp_pos();
+
+ auto * inp_attn = build_attn_inp_kv_unified();
+
+ for (int il = 0; il < n_layer; ++il) {
+ ggml_tensor * inpSA = inpL;
+
+ // norm
+ {
+ cur = build_norm(inpL,
+ model.layers[il].attn_norm, NULL,
+ LLM_NORM_RMS, il);
+ cb(cur, "attn_norm", il);
+ }
+
+ // self-attention
+ {
+ ggml_tensor * Qcur = build_lora_mm(model.layers[il].wq, cur);
+ cb(Qcur, "Qcur", il);
+ if (model.layers[il].bq) {
+ Qcur = ggml_add(ctx0, Qcur, model.layers[il].bq);
+ cb(Qcur, "Qcur", il);
+ }
+
+ ggml_tensor * Kcur = build_lora_mm(model.layers[il].wk, cur);
+ cb(Kcur, "Kcur", il);
+ if (model.layers[il].bk) {
+ Kcur = ggml_add(ctx0, Kcur, model.layers[il].bk);
+ cb(Kcur, "Kcur", il);
+ }
+
+ ggml_tensor * Vcur = build_lora_mm(model.layers[il].wv, cur);
+ cb(Vcur, "Vcur", il);
+ if (model.layers[il].bv) {
+ Vcur = ggml_add(ctx0, Vcur, model.layers[il].bv);
+ cb(Vcur, "Vcur", il);
+ }
+
+ Qcur = ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens);
+ Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens);
+ Vcur = ggml_reshape_3d(ctx0, Vcur, n_embd_head, n_head_kv, n_tokens);
+
+ Qcur = ggml_rope_ext(
+ ctx0, Qcur, inp_pos, nullptr,
+ n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+ ext_factor, attn_factor, beta_fast, beta_slow
+ );
+
+ Kcur = ggml_rope_ext(
+ ctx0, Kcur, inp_pos, nullptr,
+ n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
+ ext_factor, attn_factor, beta_fast, beta_slow
+ );
+
+ cb(Qcur, "Qcur", il);
+ cb(Kcur, "Kcur", il);
+ cb(Vcur, "Vcur", il);
+
+ cur = build_attn(inp_attn, gf,
+ model.layers[il].wo, NULL,
+ Qcur, Kcur, Vcur, nullptr, nullptr, 1.0f/sqrtf(float(n_embd_head)), il);
+ }
+
+ if (il == n_layer - 1) {
+ // skip computing output for unused tokens
+ ggml_tensor * inp_out_ids = build_inp_out_ids();
+ cur = ggml_get_rows(ctx0, cur, inp_out_ids);
+ inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
+ }
+
+ ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
+ cb(ffn_inp, "ffn_inp", il);
+
+ // feed-forward network
+ {
+ cur = build_norm(ffn_inp,
+ model.layers[il].ffn_norm, NULL,
+ LLM_NORM_RMS, il);
+ cb(cur, "ffn_norm", il);
+
+ cur = build_ffn(cur,
+ model.layers[il].ffn_up, NULL, NULL,
+ model.layers[il].ffn_gate, NULL, NULL,
+ model.layers[il].ffn_down, NULL, NULL,
+ NULL,
+ LLM_FFN_SILU, LLM_FFN_PAR, il);
+ cb(cur, "ffn_out", il);
+ }
+
+ cur = ggml_add(ctx0, cur, ffn_inp);
+
+ cur = build_cvec(cur, il);
+ cb(cur, "l_out", il);
+
+ // input for next layer
+ inpL = cur;
+ }
+
+ cur = inpL;
+
+ cur = build_norm(cur,
+ model.output_norm, NULL,
+ LLM_NORM_RMS, -1);
+
+ cb(cur, "result_norm", -1);
+ res->t_embd = cur;
+
+ // lm_head
+ cur = build_lora_mm(model.output, cur);
+
+ cb(cur, "result_output", -1);
+ res->t_logits = cur;
+
+ ggml_build_forward_expand(gf, cur);
+ }
+};
+
struct llm_build_arcee : public llm_graph_context {
llm_build_arcee(const llama_model & model, const llm_graph_params & params, ggml_cgraph * gf) : llm_graph_context(params) {
const int64_t n_embd_head = hparams.n_embd_head_v;
{
llm = std::make_unique<llm_build_arcee>(*this, params, gf);
} break;
+ case LLM_ARCH_ERNIE4_5:
+ {
+ llm = std::make_unique<llm_build_ernie4_5>(*this, params, gf);
+ } break;
default:
GGML_ABORT("fatal error");
}
case LLM_ARCH_BAILINGMOE:
case LLM_ARCH_NEO_BERT:
case LLM_ARCH_ARCEE:
+ case LLM_ARCH_ERNIE4_5:
return LLAMA_ROPE_TYPE_NORM;
// the pairs of head values are offset by n_rot/2
overview / pkg / ggml / sources / llama.cpp / commitdiff