Hugging Face's logo

CrashOverrideX
/
Quillan-Ronin

Text Generation
Transformers
Safetensors
GGUF
English
llama
causal-lm
multimodal
image-generation
video-generation
audio-generation
text-generation-inference
Model card Files
xet
Community
Quillan-Ronin
File size: 4,187 Bytes
e737d04
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
138
139
140
141
142
143
144
145
146
147
148
149
 
#include "backend-dispatched.h"
#include "backend-virgl-apir.h"
#include "ggml-backend-impl.h"
#include "ggml-backend.h"
#include "ggml-impl.h"

#include <cstdint>

uint32_t backend_device_get_device_count(apir_encoder * enc, apir_decoder * dec, virgl_apir_context * ctx) {
    GGML_UNUSED(ctx);
    GGML_UNUSED(ctx);
    GGML_UNUSED(dec);

    int32_t dev_count = reg->iface.get_device_count(reg);
    apir_encode_int32_t(enc, &dev_count);

    return 0;
}

uint32_t backend_device_get_count(apir_encoder * enc, apir_decoder * dec, virgl_apir_context * ctx) {
    GGML_UNUSED(ctx);
    GGML_UNUSED(ctx);
    GGML_UNUSED(dec);

    int32_t dev_count = reg->iface.get_device_count(reg);
    apir_encode_int32_t(enc, &dev_count);

    return 0;
}

uint32_t backend_device_get_name(apir_encoder * enc, apir_decoder * dec, virgl_apir_context * ctx) {
    GGML_UNUSED(ctx);
    GGML_UNUSED(dec);

    const char * string = dev->iface.get_name(dev);

    const size_t string_size = strlen(string) + 1;
    apir_encode_array_size(enc, string_size);
    apir_encode_char_array(enc, string, string_size);

    return 0;
}

uint32_t backend_device_get_description(apir_encoder * enc, apir_decoder * dec, virgl_apir_context * ctx) {
    GGML_UNUSED(ctx);
    GGML_UNUSED(dec);

    const char * string = dev->iface.get_description(dev);

    const size_t string_size = strlen(string) + 1;
    apir_encode_array_size(enc, string_size);
    apir_encode_char_array(enc, string, string_size);

    return 0;
}

uint32_t backend_device_get_type(apir_encoder * enc, apir_decoder * dec, virgl_apir_context * ctx) {
    GGML_UNUSED(ctx);
    GGML_UNUSED(dec);

    uint32_t type = dev->iface.get_type(dev);
    apir_encode_uint32_t(enc, &type);

    return 0;
}

uint32_t backend_device_get_memory(apir_encoder * enc, apir_decoder * dec, virgl_apir_context * ctx) {
    GGML_UNUSED(ctx);
    GGML_UNUSED(dec);

    size_t free, total;
    dev->iface.get_memory(dev, &free, &total);

    apir_encode_size_t(enc, &free);
    apir_encode_size_t(enc, &total);

    return 0;
}

uint32_t backend_device_supports_op(apir_encoder * enc, apir_decoder * dec, virgl_apir_context * ctx) {
    GGML_UNUSED(ctx);

    const ggml_tensor * op = apir_decode_ggml_tensor_inplace(dec);

    bool supports_op = dev->iface.supports_op(dev, op);

    apir_encode_bool_t(enc, &supports_op);

    return 0;
}

uint32_t backend_device_get_buffer_type(apir_encoder * enc, apir_decoder * dec, virgl_apir_context * ctx) {
    GGML_UNUSED(ctx);
    GGML_UNUSED(dec);

    ggml_backend_buffer_type_t bufft = dev->iface.get_buffer_type(dev);

    apir_encode_ggml_buffer_type(enc, bufft);

    return 0;
}

uint32_t backend_device_get_props(apir_encoder * enc, apir_decoder * dec, virgl_apir_context * ctx) {
    GGML_UNUSED(ctx);
    GGML_UNUSED(dec);

    ggml_backend_dev_props props;
    dev->iface.get_props(dev, &props);

    apir_encode_bool_t(enc, &props.caps.async);
    apir_encode_bool_t(enc, &props.caps.host_buffer);
    apir_encode_bool_t(enc, &props.caps.buffer_from_host_ptr);
    apir_encode_bool_t(enc, &props.caps.events);

    return 0;
}

uint32_t backend_device_buffer_from_ptr(apir_encoder * enc, apir_decoder * dec, virgl_apir_context * ctx) {
    GGML_UNUSED(ctx);
    GGML_UNUSED(dec);

    uint32_t shmem_res_id;
    apir_decode_virtgpu_shmem_res_id(dec, &shmem_res_id);

    void * shmem_ptr = ctx->iface->get_shmem_ptr(ctx->ctx_id, shmem_res_id);
    if (!shmem_ptr) {
        GGML_LOG_ERROR(GGML_VIRTGPU_BCK "%s: Couldn't get the shmem addr from virgl\n", __func__);
        apir_decoder_set_fatal(dec);
        return 1;
    }

    size_t size;
    apir_decode_size_t(dec, &size);
    size_t max_tensor_size;
    apir_decode_size_t(dec, &max_tensor_size);

    ggml_backend_buffer_t buffer;
    buffer = dev->iface.buffer_from_host_ptr(dev, shmem_ptr, size, max_tensor_size);

    apir_encode_ggml_buffer(enc, buffer);
    apir_encode_ggml_buffer_type(enc, buffer->buft);

    if (buffer) {
        apir_track_backend_buffer(buffer);
    }

    return 0;
}