npm - @novastera-oss/llamarn - Versions diffs - 0.4.1 → 0.4.3-beta4 - Mend

@novastera-oss/llamarn 0.4.1 → 0.4.3-beta4

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (976) hide show

package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.comp DELETED Viewed

@@ -1,154 +0,0 @@
-#extension GL_EXT_control_flow_attributes : enable
-#extension GL_EXT_shader_16bit_storage : require
-#extension GL_EXT_shader_8bit_storage : require
-#if USE_SUBGROUP_ADD
-#extension GL_KHR_shader_subgroup_basic : require
-#extension GL_KHR_shader_subgroup_arithmetic : require
-#endif
-#ifdef MUL_MAT_ID
-#define EXPERT_COUNT 8
-#endif
-#include "types.comp"
-layout (binding = 0) readonly buffer A {A_TYPE data_a[];};
-layout (binding = 1) readonly buffer B {B_TYPE data_b[];};
-layout (binding = 1) readonly buffer BV2 {B_TYPE_VEC2 data_b_v2[];};
-layout (binding = 1) readonly buffer BV4 {B_TYPE_VEC4 data_b_v4[];};
-layout (binding = 2) writeonly buffer D {D_TYPE data_d[];};
-#ifdef MUL_MAT_ID
-layout (binding = 3) readonly buffer IDS {int data_ids[];};
-#endif
-#include "dequant_funcs.comp"
-layout (push_constant) uniform parameter
-{
-    uint ncols;
-    uint stride_a;
-    uint stride_b;
-    uint stride_d;
-    uint batch_stride_a;
-    uint batch_stride_b;
-    uint batch_stride_d;
-#ifdef MUL_MAT_ID
-    uint nei0;
-    uint ne11;
-#else
-    uint ne02;
-    uint ne12;
-    uint broadcast2;
-    uint broadcast3;
-#endif
-} p;
-void get_offsets(out uint a_offset, out uint b_offset, out uint d_offset) {
-#ifdef MUL_MAT_ID
-    const uint expert_idx = gl_GlobalInvocationID.y;
-#else
-    const uint batch_idx = gl_GlobalInvocationID.y;
-#endif
-#ifndef MUL_MAT_ID
-    uint batch_idx_a = 0;
-    if (batch_idx != 0) {
-        const uint i13 = batch_idx / p.ne12;
-        const uint i12 = batch_idx % p.ne12;
-        const uint i03 = i13 / p.broadcast3;
-        const uint i02 = i12 / p.broadcast2;
-        batch_idx_a = i03 * p.ne02 + i02;
-    }
-#else
-    const uint expert_id = data_ids[expert_idx];
-#endif
-    a_offset =
-#ifdef MUL_MAT_ID
-            expert_id * p.batch_stride_a;
-#else
-            batch_idx_a * p.batch_stride_a;
-#endif
-    b_offset =
-#ifdef MUL_MAT_ID
-            (expert_idx % p.ne11) * p.stride_b;
-#else
-            batch_idx * p.batch_stride_b;
-#endif
-    d_offset =
-#ifdef MUL_MAT_ID
-            expert_idx * p.stride_d;
-#else
-            batch_idx * p.batch_stride_d;
-#endif
-}
-layout (constant_id = 0) const uint BLOCK_SIZE = 32;
-layout (constant_id = 1) const uint NUM_ROWS = 1;
-layout (constant_id = 2) const uint NUM_COLS = 1;
-shared FLOAT_TYPE tmpsh[NUM_COLS][NUM_ROWS][BLOCK_SIZE];
-void reduce_result(FLOAT_TYPE temp[NUM_COLS][NUM_ROWS], const in uint32_t d_offset, const in uint32_t first_row, const in uint32_t num_rows, const in uint32_t tid) {
-    // subgroupAdd is probably faster on devices that support it,
-    // particularly when the workgroup has more than one subgroup
-#if USE_SUBGROUP_ADD
-    // sum up partial sums within a subgroup
-    [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-        [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-            temp[j][n] = subgroupAdd(temp[j][n]);
-        }
-    }
-    // Go through shared memory to sum partials across subgroups
-    if (gl_SubgroupInvocationID == 0) {
-        [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-            [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-                tmpsh[j][n][gl_SubgroupID] = temp[j][n];
-            }
-        }
-    }
-    barrier();
-    if (tid == 0) {
-        [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-            [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-                temp[j][n] = FLOAT_TYPE(0);
-                [[unroll]] for (uint s = 0; s < gl_NumSubgroups; ++s) {
-                    temp[j][n] += tmpsh[j][n][s];
-                }
-                data_d[j*p.batch_stride_d + d_offset + first_row + n] = D_TYPE(temp[j][n]);
-            }
-        }
-    }
-#else
-    // sum up partial sums and write back result
-    [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-        [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-            tmpsh[j][n][tid] = temp[j][n];
-        }
-    }
-    barrier();
-    [[unroll]] for (uint s = BLOCK_SIZE/2; s > 0; s >>= 1) {
-        if (tid < s) {
-            [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-                [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-                    tmpsh[j][n][tid] += tmpsh[j][n][tid + s];
-                }
-            }
-        }
-        barrier();
-    }
-    if (tid == 0) {
-        [[unroll]] for (uint j = 0; j < NUM_COLS; ++j) {
-            [[unroll]] for (uint n = 0; n < num_rows; ++n) {
-                data_d[j*p.batch_stride_d + d_offset + first_row + n] = D_TYPE(tmpsh[j][n][0]);
-            }
-        }
-    }
-#endif
-}

package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.comp DELETED Viewed

@@ -1,105 +0,0 @@
-#extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
-#extension GL_EXT_shader_explicit_arithmetic_types_int16 : require
-#extension GL_EXT_shader_explicit_arithmetic_types_int8 : require
-#include "types.comp"
-// Each iqs value maps to a 32-bit integer
-#if defined(DATA_A_Q4_0)
-i32vec2 repack(uint ib, uint iqs) {
-    // Use 2-byte loads since a q4_0 block (18 bytes) is not divisible by 4
-    const u16vec2 quants = u16vec2(data_a[ib].qs[iqs * 2    ],
-                                   data_a[ib].qs[iqs * 2 + 1]);
-    const uint32_t vui = pack32(quants);
-    return i32vec2( vui       & 0x0F0F0F0F,
-                   (vui >> 4) & 0x0F0F0F0F);
-}
-ACC_TYPE mul_q8_1(int32_t q_sum, float da, vec2 dsb) {
-    return ACC_TYPE(da * (float(q_sum) * dsb.x - 8.0f * dsb.y));
-}
-#endif
-#if defined(DATA_A_Q4_1)
-i32vec2 repack(uint ib, uint iqs) {
-    // Use 4-byte loads since a q4_1 block (20 bytes) is divisible by 4
-    const uint32_t vui = data_a_packed32[ib].qs[iqs];
-    return i32vec2( vui       & 0x0F0F0F0F,
-                   (vui >> 4) & 0x0F0F0F0F);
-}
-ACC_TYPE mul_q8_1(int32_t q_sum, vec2 dma, vec2 dsb) {
-    return ACC_TYPE(float(q_sum) * dma.x * dsb.x + dma.y * dsb.y);
-}
-#endif
-#if defined(DATA_A_Q5_0)
-i32vec2 repack(uint ib, uint iqs) {
-    // Use 2-byte loads since a q5_0 block (22 bytes) is not divisible by 4
-    const u16vec2 quants = u16vec2(data_a[ib].qs[iqs * 2    ],
-                                   data_a[ib].qs[iqs * 2 + 1]);
-    const uint32_t vui = pack32(quants);
-    const int32_t qh = int32_t((uint32_t(data_a[ib].qh[1]) << 16 | data_a[ib].qh[0]) >> (4 * iqs));
-    const int32_t v0 = int32_t(vui & 0x0F0F0F0F)
-                     | ((qh & 0xF) * 0x02040810) & 0x10101010; // (0,1,2,3) -> (4,12,20,28)
-    const int32_t v1 = int32_t((vui >> 4) & 0x0F0F0F0F)
-                     | (((qh >> 16) & 0xF) * 0x02040810) & 0x10101010; // (16,17,18,19) -> (4,12,20,28)
-    return i32vec2(v0, v1);
-}
-ACC_TYPE mul_q8_1(int32_t q_sum, float da, vec2 dsb) {
-    return ACC_TYPE(da * (float(q_sum) * dsb.x - 16.0f * dsb.y));
-}
-#endif
-#if defined(DATA_A_Q5_1)
-i32vec2 repack(uint ib, uint iqs) {
-    // Use 4-byte loads since a q5_1 block (24 bytes) is divisible by 4
-    const uint32_t vui = data_a_packed32[ib].qs[iqs];
-    const int32_t qh = int32_t(data_a_packed32[ib].qh >> (4 * iqs));
-    const int32_t v0 = int32_t(vui & 0x0F0F0F0F)
-                     | ((qh & 0xF) * 0x02040810) & 0x10101010; // (0,1,2,3) -> (4,12,20,28)
-    const int32_t v1 = int32_t((vui >> 4) & 0x0F0F0F0F)
-                     | (((qh >> 16) & 0xF) * 0x02040810) & 0x10101010; // (16,17,18,19) -> (4,12,20,28)
-    return i32vec2(v0, v1);
-}
-ACC_TYPE mul_q8_1(int32_t q_sum, vec2 dma, vec2 dsb) {
-    return ACC_TYPE(float(q_sum) * dma.x * dsb.x + dma.y * dsb.y);
-}
-#endif
-#if defined(DATA_A_Q8_0)
-int32_t repack(uint ib, uint iqs) {
-    // Use 2-byte loads since a q8_0 block (34 bytes) is not divisible by 4
-    return pack32(i16vec2(data_a[ib].qs[iqs * 2    ],
-                          data_a[ib].qs[iqs * 2 + 1]));
-}
-ACC_TYPE mul_q8_1(int32_t q_sum, float da, vec2 dsb) {
-    return ACC_TYPE(float(q_sum) * da * dsb.x);
-}
-#endif
-#if defined(DATA_A_Q4_0) || defined(DATA_A_Q5_0) || defined(DATA_A_Q8_0) || defined(DATA_A_IQ1_S) || defined(DATA_A_IQ2_XXS) || defined(DATA_A_IQ2_XS) || defined(DATA_A_IQ2_S) || defined(DATA_A_IQ3_XXS) || defined(DATA_A_IQ3_S) || defined(DATA_A_IQ4_XS) || defined(DATA_A_IQ4_NL)
-FLOAT_TYPE get_d(uint ib) {
-    return FLOAT_TYPE(data_a[ib].d);
-}
-#endif
-#if defined(DATA_A_MXFP4)
-FLOAT_TYPE get_d(uint ib) {
-    return FLOAT_TYPE(e8m0_to_fp32(data_a[ib].e));
-}
-#endif
-#if defined(DATA_A_Q4_1) || defined(DATA_A_Q5_1)
-FLOAT_TYPE_VEC2 get_dm(uint ib) {
-    return FLOAT_TYPE_VEC2(data_a_packed32[ib].dm);
-}
-#endif

package/cpp/llama.cpp/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.comp DELETED Viewed

@@ -1,55 +0,0 @@
-#include "types.comp"
-#extension GL_EXT_shader_16bit_storage : require
-#include "rte.comp"
-layout(local_size_x = 1, local_size_y = 256, local_size_z = 1) in;
-layout (binding = 0) readonly buffer X {A_TYPE data_a[];};
-layout (binding = 1) readonly buffer Y {int data_pos[];};
-layout (binding = 2) readonly buffer Z {float data_ff[];};
-layout (binding = 3) writeonly buffer D {D_TYPE data_d[];};
-layout (push_constant) uniform parameter {
-    uint ncols;
-    uint n_dims;
-    float freq_scale;
-    uint p_delta_rows;
-    float freq_base;
-    float ext_factor;
-    float attn_factor;
-    float corr_dims[2];
-    float theta_scale;
-    uint has_ff;
-    uint ne02;
-    uint s1;
-    uint s2;
-    int sections[4];
-    uint is_back;
-} p;
-float rope_yarn_ramp(const float low, const float high, const uint i0) {
-    const float y = (i0 / 2 - low) / max(0.001f, high - low);
-    return 1.0f - min(1.0f, max(0.0f, y));
-}
-void rope_yarn(const float theta_extrap, const uint i0, out float cos_theta, out float sin_theta) {
-    float mscale = p.attn_factor;
-    // Get n-d rotational scaling corrected for extrapolation
-    float theta_interp = p.freq_scale * theta_extrap;
-    float theta = theta_interp;
-    if (p.ext_factor != 0.0f) {
-        float ramp_mix = rope_yarn_ramp(p.corr_dims[0], p.corr_dims[1], i0) * p.ext_factor;
-        theta = theta_interp * (1 - ramp_mix) + theta_extrap * ramp_mix;
-        // Get n-d magnitude scaling corrected for interpolation
-        mscale *= 1.0f + 0.1f * log(1.0f / p.freq_scale);
-    }
-    // Backprogagation uses inverted rotation
-    if (p.is_back != 0) {
-        theta = -theta;
-    }
-    cos_theta = cos(theta) * mscale;
-    sin_theta = sin(theta) * mscale;
-}

package/cpp/llama.cpp/ggml/src/ggml-webgpu/wgsl-shaders/cpy.wgsl DELETED Viewed

@@ -1,60 +0,0 @@
-enable f16;
-@group(0) @binding(0)
-var<storage, read_write> src: array<f32>;
-@group(0) @binding(1)
-var<storage, read_write> dst: array<f16>;
-struct Params {
-    ne: u32,            // total number of elements
-    offset_src: u32,    // in elements
-    offset_dst: u32,    // in elements
-    // Strides (in elements) — may be permuted
-    stride_src0: u32,
-    stride_src1: u32,
-    stride_src2: u32,
-    stride_src3: u32,
-    stride_dst0: u32,
-    stride_dst1: u32,
-    stride_dst2: u32,
-    stride_dst3: u32,
-    // Logical shape (same for both tensors)
-    ne0: u32,
-    ne1: u32,
-    ne2: u32,
-    ne3: u32,
-};
-@group(0) @binding(2)
-var<uniform> params: Params;
-override wg_size: u32;
-@compute @workgroup_size(wg_size)
-fn main(@builtin(global_invocation_id) gid: vec3<u32>) {
-    if (gid.x >= params.ne) {
-        return;
-    }
-    var i = gid.x;
-    let i3 = i / (params.ne2 * params.ne1 * params.ne0);
-    i = i % (params.ne2 * params.ne1 * params.ne0);
-    let i2 = i / (params.ne1 * params.ne0);
-    i = i % (params.ne1 * params.ne0);
-    let i1 = i / params.ne0;
-    let i0 = i % params.ne0;
-    let src_idx = i0 * params.stride_src0 + i1 * params.stride_src1 +
-                  i2 * params.stride_src2 + i3 * params.stride_src3;
-    let dst_idx = i0 * params.stride_dst0 + i1 * params.stride_dst1 +
-                  i2 * params.stride_dst2 + i3 * params.stride_dst3;
-    dst[params.offset_dst + dst_idx] = f16(src[params.offset_src + src_idx]);
-}

package/cpp/llama.cpp/ggml/src/ggml-zdnn/ggml-zdnn-impl.h DELETED Viewed

@@ -1,97 +0,0 @@
-#ifndef GGML_ZDNN_IMPL
-#define GGML_ZDNN_IMPL
-#include "zdnn.h"
-#include "ggml.h"
-#include "ggml-zdnn.h"
-#include <vector>
-#include <memory>
-#include <vecintrin.h>
-#define GGML_ZDNN_NAME    "zDNN"
-#define GGML_ZDNN_VERSION ZDNN_VERNUM
-#define vec_neg(a)    (-(a))                // Vector Negate
-#define vec_add(a, b) ((a) + (b))           // Vector Add
-#define vec_sub(a, b) ((a) - (b))           // Vector Subtract
-#define vec_mul(a, b) ((a) * (b))           // Vector Multiply
-#define vec_div(a, b) ((a) / (b))           // Vector Divide
-#define vec_sl(a, b)  ((a) << (b))          // Vector Shift Left
-#define vec_sra(a, b) ((a) >> (b))          // Vector Shift Right
-#define vec_sr(a, b)  ((a) >> (b))          // Vector Shift Right Algebraic
-#define vec_slo(a, b) vec_slb(a, (b) << 64) // Vector Shift Left by Octet
-#define vec_sro(a, b) vec_srb(a, (b) << 64) // Vector Shift Right by Octet
-#ifndef vec_and
-#define vec_and(a, b) ((a) & (b)) // Vector AND
-#endif
-#ifndef vec_or
-#define vec_or(a, b)  ((a) | (b)) // Vector OR
-#endif
-#ifndef vec_xor
-#define vec_xor(a, b) ((a) ^ (b)) // Vector XOR
-#endif
-typedef   signed char char8x16_t  __attribute__((vector_size(16)));
-typedef unsigned char uchar8x16_t __attribute__((vector_size(16)));
-typedef int8_t   int8x16_t  __attribute__((vector_size(16)));
-typedef int16_t  int16x8_t  __attribute__((vector_size(16)));
-typedef int32_t  int32x4_t  __attribute__((vector_size(16)));
-typedef uint8_t  uint8x16_t __attribute__((vector_size(16)));
-typedef uint16_t uint16x8_t __attribute__((vector_size(16)));
-typedef uint32_t uint32x4_t __attribute__((vector_size(16)));
-typedef float float32x4_t   __attribute__((vector_size(16)));
-typedef double double64x2_t __attribute__((vector_size(16)));
-typedef   signed long long long64x2_t  __attribute__((vector_size(16)));
-typedef unsigned long long ulong64x2_t __attribute__((vector_size(16)));
-#define ZDNN_CHECK(stmt)                \
-    do {                                \
-        zdnn_status status = (stmt);    \
-        GGML_ASSERT(status == ZDNN_OK); \
-    } while (0);
-struct ggml_backend_zdnn_device_context {
-    int zdnn_device;
-    int zdnn_device_ref_count;
-    bool has_parmblkformat_0;
-    bool has_parmblkformat_1;
-    size_t max_size;
-    char name[128];
-};
-struct ggml_backend_zdnn_context {
-    int device;
-    ggml_cgraph * gf;
-};
-struct ggml_backend_zdnn_buffer {
-    void * data;
-    size_t size;
-    zdnn_tensor_desc pre_tfm_desc;
-    zdnn_tensor_desc tfm_desc;
-    zdnn_ztensor     ztensor;
-    char name[GGML_MAX_NAME];
-};
-struct ggml_backend_zdnn_buffer_context {
-    void * all_data;
-    size_t all_size;
-    bool owned;
-    int n_buffers;
-    std::vector<std::unique_ptr<ggml_backend_zdnn_buffer>> buffers;
-};
-#endif  // GGML_ZDNN_IMPL

package/cpp/llama.cpp/models/ggml-vocab-aquila.gguf DELETED Viewed

Binary file

package/cpp/llama.cpp/models/ggml-vocab-baichuan.gguf DELETED Viewed

Binary file

package/cpp/llama.cpp/models/ggml-vocab-bert-bge.gguf DELETED Viewed

Binary file

package/cpp/llama.cpp/models/ggml-vocab-bert-bge.gguf.inp DELETED Viewed

@@ -1,112 +0,0 @@
-ied 4 ½ months
-__ggml_vocab_test__
-Äpfel
-__ggml_vocab_test__
-__ggml_vocab_test__
-__ggml_vocab_test__
-__ggml_vocab_test__
-__ggml_vocab_test__
-__ggml_vocab_test__
-__ggml_vocab_test__
-__ggml_vocab_test__
-__ggml_vocab_test__
-__ggml_vocab_test__
-Hello world
-__ggml_vocab_test__
- Hello world
-__ggml_vocab_test__
-Hello World
-__ggml_vocab_test__
- Hello World
-__ggml_vocab_test__
- Hello World!
-__ggml_vocab_test__
-Hello, world!
-__ggml_vocab_test__
- Hello, world!
-__ggml_vocab_test__
- this is 🦙.cpp
-__ggml_vocab_test__
-w048 7tuijk dsdfhu
-__ggml_vocab_test__
-нещо на Български
-__ggml_vocab_test__
-កាន់តែពិសេសអាចខលចេញ
-__ggml_vocab_test__
-🚀 (normal) 😶‍🌫️ (multiple emojis concatenated) ✅ (only emoji that has its own token)
-__ggml_vocab_test__
-Hello
-__ggml_vocab_test__
- Hello
-__ggml_vocab_test__
-  Hello
-__ggml_vocab_test__
-   Hello
-__ggml_vocab_test__
-    Hello
-__ggml_vocab_test__
-    Hello
-    Hello
-__ggml_vocab_test__
- (
-__ggml_vocab_test__
- =
-__ggml_vocab_test__
-' era
-__ggml_vocab_test__
-Hello, y'all! How are you 😁 ?我想在apple工作1314151天～
-__ggml_vocab_test__
-!!!!!!
-__ggml_vocab_test__
-3
-__ggml_vocab_test__
-33
-__ggml_vocab_test__
-333
-__ggml_vocab_test__
-3333
-__ggml_vocab_test__
-33333
-__ggml_vocab_test__
-333333
-__ggml_vocab_test__
-3333333
-__ggml_vocab_test__
-33333333
-__ggml_vocab_test__
-333333333
-__ggml_vocab_test__
-Cửa Việt
-__ggml_vocab_test__
- discards
-__ggml_vocab_test__
-🚀 (normal) 😶‍🌫️ (multiple emojis concatenated) ✅ 🦙🦙 3 33 333 3333 33333 333333 3333333 33333333 3.3 3..3 3...3 កាន់តែពិសេសអាច😁 ?我想在apple工作1314151天～ ------======= нещо на Български ''''''```````""""......!!!!!!?????? I've been 'told he's there, 'RE you sure? 'M not sure I'll make it, 'D you like some tea? We'Ve a'lL
-__ggml_vocab_test__

package/cpp/llama.cpp/models/ggml-vocab-bert-bge.gguf.out DELETED Viewed

@@ -1,46 +0,0 @@
- 29464 2094 1018 1092 2706
- 9706 7959 2140
- 7592 2088
- 7592 2088
- 7592 2088
- 7592 2088
- 7592 2088 999
- 7592 1010 2088 999
- 7592 1010 2088 999
- 2023 2003 100 1012 18133 2361
- 1059 2692 18139 1021 8525 28418 2243 16233 20952 6979
- 1192 15290 29754 14150 1192 10260 1181 29755 29436 29741 10260 16856 29747 23925 10325
- 100
- 100 1006 3671 1007 100 1006 3674 7861 29147 2483 9530 16280 23854 1007 100 1006 2069 7861 29147 2072 2008 2038 2049 2219 19204 1007
- 7592
- 7592
- 7592
- 7592
- 7592
- 7592 7592
- 1006
- 1027
- 1005 3690
- 7592 1010 1061 1005 2035 999 2129 2024 2017 100 1029 1855 100 100 6207 100 100 14677 23632 22203 1811 1995
- 999 999 999 999 999 999
- 1017
- 3943
- 21211
- 21211 2509
- 21211 22394
- 21211 22394 2509
- 21211 22394 22394
- 21211 22394 22394 2509
- 21211 22394 22394 22394
- 12731 2050 19710
- 5860 18117
- 100 1006 3671 1007 100 1006 3674 7861 29147 2483 9530 16280 23854 1007 100 100 1017 3943 21211 21211 2509 21211 22394 21211 22394 2509 21211 22394 22394 21211 22394 22394 2509 1017 1012 1017 1017 1012 1012 1017 1017 1012 1012 1012 1017 100 1029 1855 100 100 6207 100 100 14677 23632 22203 1811 1995 1011 1011 1011 1011 1011 1011 1027 1027 1027 1027 1027 1027 1027 1192 15290 29754 14150 1192 10260 1181 29755 29436 29741 10260 16856 29747 23925 10325 1005 1005 1005 1005 1005 1005 1036 1036 1036 1036 1036 1036 1036 1000 1000 1000 1000 1012 1012 1012 1012 1012 1012 999 999 999 999 999 999 1029 1029 1029 1029 1029 1029 1045 1005 2310 2042 1005 2409 2002 1005 1055 2045 1010 1005 2128 2017 2469 1029 1005 1049 2025 2469 1045 1005 2222 2191 2009 1010 1005 1040 2017 2066 2070 5572 1029 2057 1005 2310 1037 1005 2222

package/cpp/llama.cpp/models/ggml-vocab-command-r.gguf DELETED Viewed

Binary file