diff --git a/ggml.c b/ggml.c index a2596db73..b96a82a41 100644 --- a/ggml.c +++ b/ggml.c @@ -13409,7 +13409,7 @@ static void ggml_compute_forward_soft_max_f32( const float slope = (max_bias > 0.0f) ? h < n_head_log2 ? powf(m0, h + 1) : powf(m1, 2*(h - n_head_log2) + 1) : 1.0f; float * sp = (float *)((char *) src0->data + i1*src0->nb[1]); - float * dp = (float *)((char *) dst->data + i1*dst->nb[1]); + float * dp = (float *)((char *) dst->data + i1*dst->nb[1]); // broadcast the mask across rows ggml_fp16_t * mp_f16 = src1 ? (ggml_fp16_t *)((char *) src1->data) + (i1%ne01)*ne00 : NULL; diff --git a/llama.cpp b/llama.cpp index eff22bb96..360d7eb90 100644 --- a/llama.cpp +++ b/llama.cpp @@ -4423,6 +4423,8 @@ static void llm_load_vocab( vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_STARCODER; } else if ( tokenizer_pre == "gpt-2" || + tokenizer_pre == "jina-es" || + tokenizer_pre == "jina-de" || tokenizer_pre == "jina-v2-es" || tokenizer_pre == "jina-v2-de") { vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_GPT2; diff --git a/unicode-data.h b/unicode-data.h index a9c2fd258..3cccf2068 100644 --- a/unicode-data.h +++ b/unicode-data.h @@ -14,4 +14,4 @@ extern const std::vector> unicode_ranges_punctuati extern const std::vector> unicode_ranges_symbol; extern const std::vector> unicode_ranges_control; extern const std::multimap unicode_map_nfd; -extern const std::map unicode_map_lowercase; \ No newline at end of file +extern const std::map unicode_map_lowercase;