models : move the token embedding norms to the first layer (#20943)
* models : move the token embedding norms to the first layer * cont : fix LLM_TENSOR_CONV1D + fix il indexing
This commit is contained in:
+2
-2
@@ -2564,7 +2564,7 @@ static const std::map<llm_tensor, llm_tensor_info> LLM_TENSOR_INFOS = {
|
||||
{LLM_TENSOR_TOKEN_EMBD, {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}},
|
||||
{LLM_TENSOR_POS_EMBD, {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}},
|
||||
{LLM_TENSOR_TOKEN_TYPES, {LLM_TENSOR_LAYER_INPUT, GGML_OP_GET_ROWS}},
|
||||
{LLM_TENSOR_TOKEN_EMBD_NORM, {LLM_TENSOR_LAYER_INPUT, GGML_OP_MUL}},
|
||||
{LLM_TENSOR_TOKEN_EMBD_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}}, // do the norms on the first layer (not the input layer)
|
||||
{LLM_TENSOR_OUTPUT, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}},
|
||||
{LLM_TENSOR_CLS, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}},
|
||||
{LLM_TENSOR_CLS_OUT, {LLM_TENSOR_LAYER_OUTPUT, GGML_OP_MUL_MAT}},
|
||||
@@ -2725,7 +2725,7 @@ static const std::map<llm_tensor, llm_tensor_info> LLM_TENSOR_INFOS = {
|
||||
{LLM_TENSOR_LAUREL_POST_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
|
||||
// this tensor is loaded for T5, but never used
|
||||
{LLM_TENSOR_DEC_CROSS_ATTN_REL_B, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_NONE}},
|
||||
{LLM_TENSOR_CONV1D, {LLM_TENSOR_LAYER_INPUT, GGML_OP_IM2COL}},
|
||||
{LLM_TENSOR_CONV1D, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_IM2COL}},
|
||||
{LLM_TENSOR_POS_NET_NORM, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
|
||||
{LLM_TENSOR_POS_NET_NORM1, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
|
||||
{LLM_TENSOR_POS_NET_NORM2, {LLM_TENSOR_LAYER_REPEATING, GGML_OP_MUL}},
|
||||
|
||||
Reference in New Issue
Block a user