Hugging Face's logo

tda45
/
TdAI

Safetensors
GGUF
Turkish
llama
Llama-3
instruct
finetune
chatml
gpt4
synthetic data
distillation
function calling
json mode
axolotl
roleplaying
chat
Model card Files
xet
 Community
TdAI
File size: 4,441 Bytes
8efb28e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
#pragma once

#include "ggml.h"
#include "ggml-cpp.h"
#include "clip.h"
#include "clip-impl.h"
#include "clip-model.h"

#include <vector>
#include <functional>

#define DEFAULT_INTERPOLATION_MODE (GGML_SCALE_MODE_BILINEAR | GGML_SCALE_FLAG_ANTIALIAS)

struct build_vit_opts {
    ggml_tensor * attn_mask = nullptr;
};

struct clip_graph {
    const clip_model & model;
    const clip_hparams & hparams;
    projector_type proj_type;

    // we only support single image per batch
    const clip_image_f32 & img;

    const int patch_size;
    const int n_patches_x;
    const int n_patches_y;
    const int n_patches;
    const int n_embd;
    const int n_head;
    const int n_head_kv;
    const int d_head;
    const int n_layer;
    const int n_mmproj_embd;
    const float eps;
    float kq_scale; // TODO: maybe move this to hparams
    const clip_flash_attn_type flash_attn_type;

    // TODO [QWEN_VIDEO]: improve this in the future
    int n_batch = 1;

    ggml_context_ptr ctx0_ptr;
    ggml_context * ctx0;
    ggml_cgraph * gf;

    clip_graph(clip_ctx * ctx, const clip_image_f32 & img);

    virtual ~clip_graph() = default;
    virtual ggml_cgraph * build() = 0;

    // wrapper around ggml_mul_mat, allow hooking (e.g. LoRA, clamping) depending on the model
    // tensor w should be the weight matrix, and tensor x should be the input
    virtual ggml_tensor * build_mm(ggml_tensor * w, ggml_tensor * x) const;
    // TODO: build_mm(w, b, x) to support bias

    virtual bool support_batch() const {
        return false;
    }

    //
    // utility functions
    //
    void cb(ggml_tensor * cur0, const char * name, int il) const;

    // siglip2 naflex
    ggml_tensor * resize_position_embeddings(uint32_t interpolation_mode = DEFAULT_INTERPOLATION_MODE);

    // build vision transformer (ViT) cgraph
    // this function should cover most of the models
    // if your model has specific features, you should probably duplicate this function
    ggml_tensor * build_vit(
                ggml_tensor * inp,
                int64_t n_pos,
                norm_type norm_t,
                ffn_op_type ffn_t,
                ggml_tensor * learned_pos_embd,
                std::function<ggml_tensor *(ggml_tensor *, const clip_layer &)> add_pos,
                const build_vit_opts & opts = {});

    // build the input after conv2d (inp_raw --> patches)
    // returns tensor with shape [n_embd, n_patches]
    ggml_tensor * build_inp();

    ggml_tensor * build_inp_raw(int channels = 3);

    ggml_tensor * build_norm(
            ggml_tensor * cur,
            ggml_tensor * mw,
            ggml_tensor * mb,
            norm_type type,
            float norm_eps,
            int il) const;

    ggml_tensor * build_ffn(
            ggml_tensor * cur,
            ggml_tensor * up,
            ggml_tensor * up_b,
            ggml_tensor * gate,
            ggml_tensor * gate_b,
            ggml_tensor * down,
            ggml_tensor * down_b,
            ffn_op_type type_op,
            int il) const;

    ggml_tensor * build_attn(
            ggml_tensor * wo,
            ggml_tensor * wo_b,
            ggml_tensor * q_cur,
            ggml_tensor * k_cur,
            ggml_tensor * v_cur,
            ggml_tensor * kq_mask,
            float kq_scale,
            int il,
            ggml_tensor * sinks = nullptr) const;

    // implementation of the 2D RoPE without adding a new op in ggml
    // this is not efficient (use double the memory), but works on all backends
    // TODO: there was a more efficient which relies on ggml_view and ggml_rope_ext_inplace, but the rope inplace does not work well with non-contiguous tensors ; we should fix that and revert back to the original implementation in https://github.com/ggml-org/llama.cpp/pull/13065
    ggml_tensor * build_rope_2d(
        ggml_context * ctx0,
        ggml_tensor * cur,
        ggml_tensor * pos_a, // first half
        ggml_tensor * pos_b, // second half
        const float freq_base,
        const bool interleave_freq
    );

    // aka pixel_shuffle / pixel_unshuffle / patch_merger (Kimi-VL)
    // support dynamic resolution
    ggml_tensor * build_patch_merge_permute(ggml_tensor * cur, int scale_factor);

    // Generic function to stack frames for audio processing
    // Abstracts out the StackAudioFrames logic used by ultravox
    ggml_tensor * build_stack(ggml_tensor * cur, int32_t stack_factor, int32_t n_embed);
};