more tinygrad examples

docs/_source/_static/_output/tinygrad/compile_tensorflow.py.txt

@@ -0,0 +1,242 @@
+2024-02-06 11:09:02.997488: I external/local_tsl/tsl/cuda/cudart_stub.cc:31] Could not find cuda drivers on your machine, GPU will not be used.
+2024-02-06 11:09:03.035957: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
+2024-02-06 11:09:03.036006: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
+2024-02-06 11:09:03.036946: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
+2024-02-06 11:09:03.042522: I external/local_tsl/tsl/cuda/cudart_stub.cc:31] Could not find cuda drivers on your machine, GPU will not be used.
+2024-02-06 11:09:03.042694: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
+To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
+2024-02-06 11:09:03.830560: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT
+2024-02-06 11:09:05.325306: I tensorflow/core/grappler/devices.cc:66] Number of eligible GPUs (core count >= 8, compute capability >= 0.0): 0
+2024-02-06 11:09:05.325453: I tensorflow/core/grappler/clusters/single_machine.cc:361] Starting new session
+2024-02-06 11:09:05.420786: I tensorflow/core/grappler/devices.cc:66] Number of eligible GPUs (core count >= 8, compute capability >= 0.0): 0
+2024-02-06 11:09:05.421124: I tensorflow/core/grappler/clusters/single_machine.cc:361] Starting new session
+tinygrad: [0.270704448223114, 0.6882184743881226, 0.8074522614479065, 0.5307921767234802]
+compiled: [0.270704, 0.688218, 0.807452, 0.530792]
+keras:    [0.2707044 0.6882185 0.8074523 0.5307921]
+#include <string.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <math.h>
+#define max(x,y) ((x>y)?x:y)
+#define int64 long
+#define half __fp16
+#define uchar unsigned char
+#include <stdbool.h>
+
+float buf_0[64];
+float input0[128];
+float buf_1[2048];
+float buf_2[64];
+float buf_3[128];
+float buf_4[2048];
+float output0[16];
+float buf_5[512];
+void r_16_32(float* restrict data0, const float* restrict data1, const float* restrict data2, const float* restrict data3) {
+  float val0 = data1[0];
+  float val1 = data1[1];
+  float val2 = data1[2];
+  float val3 = data1[3];
+  float val4 = data1[4];
+  float val5 = data1[5];
+  float val6 = data1[6];
+  float val7 = data1[7];
+  float val8 = data1[8];
+  float val9 = data1[9];
+  float val10 = data1[10];
+  float val11 = data1[11];
+  float val12 = data1[12];
+  float val13 = data1[13];
+  float val14 = data1[14];
+  float val15 = data1[15];
+  float val16 = data1[16];
+  float val17 = data1[17];
+  float val18 = data1[18];
+  float val19 = data1[19];
+  float val20 = data1[20];
+  float val21 = data1[21];
+  float val22 = data1[22];
+  float val23 = data1[23];
+  float val24 = data1[24];
+  float val25 = data1[25];
+  float val26 = data1[26];
+  float val27 = data1[27];
+  float val28 = data1[28];
+  float val29 = data1[29];
+  float val30 = data1[30];
+  float val31 = data1[31];
+  for (int ridx0 = 0; ridx0 < 16; ridx0++) {
+    float acc0 = 0.0f;
+    float val32 = data2[ridx0];
+    float val33 = data2[ridx0+16];
+    float val34 = data2[ridx0+32];
+    float val35 = data2[ridx0+48];
+    float val36 = data2[ridx0+64];
+    float val37 = data2[ridx0+80];
+    float val38 = data2[ridx0+96];
+    float val39 = data2[ridx0+112];
+    float val40 = data2[ridx0+128];
+    float val41 = data2[ridx0+144];
+    float val42 = data2[ridx0+160];
+    float val43 = data2[ridx0+176];
+    float val44 = data2[ridx0+192];
+    float val45 = data2[ridx0+208];
+    float val46 = data2[ridx0+224];
+    float val47 = data2[ridx0+240];
+    float val48 = data2[ridx0+256];
+    float val49 = data2[ridx0+272];
+    float val50 = data2[ridx0+288];
+    float val51 = data2[ridx0+304];
+    float val52 = data2[ridx0+320];
+    float val53 = data2[ridx0+336];
+    float val54 = data2[ridx0+352];
+    float val55 = data2[ridx0+368];
+    float val56 = data2[ridx0+384];
+    float val57 = data2[ridx0+400];
+    float val58 = data2[ridx0+416];
+    float val59 = data2[ridx0+432];
+    float val60 = data2[ridx0+448];
+    float val61 = data2[ridx0+464];
+    float val62 = data2[ridx0+480];
+    float val63 = data2[ridx0+496];
+    float val64 = data3[ridx0];
+    float alu0 = max(((val31*val63)+((val30*val62)+((val29*val61)+((val28*val60)+((val27*val59)+((val26*val58)+((val25*val57)+((val24*val56)+((val23*val55)+((val22*val54)+((val21*val53)+((val20*val52)+((val19*val51)+((val18*val50)+((val17*val49)+((val16*val48)+((val15*val47)+((val14*val46)+((val13*val45)+((val12*val44)+((val11*val43)+((val10*val42)+((val9*val41)+((val8*val40)+((val7*val39)+((val6*val38)+((val5*val37)+((val4*val36)+((val3*val35)+((val2*val34)+((val1*val33)+((val0*val32)+acc0)))))))))))))))))))))))))))))))),0.0f);
+    data0[ridx0] = (alu0*val64);
+  }
+}
+void r_32_16(float* restrict data0, const float* restrict data1, const float* restrict data2) {
+  float val0 = data1[0];
+  float val1 = data1[1];
+  float val2 = data1[2];
+  float val3 = data1[3];
+  float val4 = data1[4];
+  float val5 = data1[5];
+  float val6 = data1[6];
+  float val7 = data1[7];
+  float val8 = data1[8];
+  float val9 = data1[9];
+  float val10 = data1[10];
+  float val11 = data1[11];
+  float val12 = data1[12];
+  float val13 = data1[13];
+  float val14 = data1[14];
+  float val15 = data1[15];
+  for (int ridx0 = 0; ridx0 < 32; ridx0++) {
+    float acc0 = 0.0f;
+    float val16 = data2[ridx0];
+    float val17 = data2[ridx0+32];
+    float val18 = data2[ridx0+64];
+    float val19 = data2[ridx0+96];
+    float val20 = data2[ridx0+128];
+    float val21 = data2[ridx0+160];
+    float val22 = data2[ridx0+192];
+    float val23 = data2[ridx0+224];
+    float val24 = data2[ridx0+256];
+    float val25 = data2[ridx0+288];
+    float val26 = data2[ridx0+320];
+    float val27 = data2[ridx0+352];
+    float val28 = data2[ridx0+384];
+    float val29 = data2[ridx0+416];
+    float val30 = data2[ridx0+448];
+    float val31 = data2[ridx0+480];
+    float alu0 = max(((val15*val31)+((val14*val30)+((val13*val29)+((val12*val28)+((val11*val27)+((val10*val26)+((val9*val25)+((val8*val24)+((val7*val23)+((val6*val22)+((val5*val21)+((val4*val20)+((val3*val19)+((val2*val18)+((val1*val17)+((val0*val16)+acc0)))))))))))))))),0.0f);
+    data0[ridx0] = alu0;
+  }
+}
+void r_4_32(float* restrict data0, const float* restrict data1, const float* restrict data2) {
+  float val0 = data1[0];
+  float val1 = data1[1];
+  float val2 = data1[2];
+  float val3 = data1[3];
+  float val4 = data1[4];
+  float val5 = data1[5];
+  float val6 = data1[6];
+  float val7 = data1[7];
+  float val8 = data1[8];
+  float val9 = data1[9];
+  float val10 = data1[10];
+  float val11 = data1[11];
+  float val12 = data1[12];
+  float val13 = data1[13];
+  float val14 = data1[14];
+  float val15 = data1[15];
+  float val16 = data1[16];
+  float val17 = data1[17];
+  float val18 = data1[18];
+  float val19 = data1[19];
+  float val20 = data1[20];
+  float val21 = data1[21];
+  float val22 = data1[22];
+  float val23 = data1[23];
+  float val24 = data1[24];
+  float val25 = data1[25];
+  float val26 = data1[26];
+  float val27 = data1[27];
+  float val28 = data1[28];
+  float val29 = data1[29];
+  float val30 = data1[30];
+  float val31 = data1[31];
+  for (int ridx0 = 0; ridx0 < 4; ridx0++) {
+    float acc0 = 0.0f;
+    float val32 = data2[ridx0];
+    float val33 = data2[ridx0+4];
+    float val34 = data2[ridx0+8];
+    float val35 = data2[ridx0+12];
+    float val36 = data2[ridx0+16];
+    float val37 = data2[ridx0+20];
+    float val38 = data2[ridx0+24];
+    float val39 = data2[ridx0+28];
+    float val40 = data2[ridx0+32];
+    float val41 = data2[ridx0+36];
+    float val42 = data2[ridx0+40];
+    float val43 = data2[ridx0+44];
+    float val44 = data2[ridx0+48];
+    float val45 = data2[ridx0+52];
+    float val46 = data2[ridx0+56];
+    float val47 = data2[ridx0+60];
+    float val48 = data2[ridx0+64];
+    float val49 = data2[ridx0+68];
+    float val50 = data2[ridx0+72];
+    float val51 = data2[ridx0+76];
+    float val52 = data2[ridx0+80];
+    float val53 = data2[ridx0+84];
+    float val54 = data2[ridx0+88];
+    float val55 = data2[ridx0+92];
+    float val56 = data2[ridx0+96];
+    float val57 = data2[ridx0+100];
+    float val58 = data2[ridx0+104];
+    float val59 = data2[ridx0+108];
+    float val60 = data2[ridx0+112];
+    float val61 = data2[ridx0+116];
+    float val62 = data2[ridx0+120];
+    float val63 = data2[ridx0+124];
+    data0[ridx0] = (1.0f/(1.0f+exp2((((val31*val63)+((val30*val62)+((val29*val61)+((val28*val60)+((val27*val59)+((val26*val58)+((val25*val57)+((val24*val56)+((val23*val55)+((val22*val54)+((val21*val53)+((val20*val52)+((val19*val51)+((val18*val50)+((val17*val49)+((val16*val48)+((val15*val47)+((val14*val46)+((val13*val45)+((val12*val44)+((val11*val43)+((val10*val42)+((val9*val41)+((val8*val40)+((val7*val39)+((val6*val38)+((val5*val37)+((val4*val36)+((val3*val35)+((val2*val34)+((val1*val33)+((val0*val32)+acc0))))))))))))))))))))))))))))))))*(-1.4426950408889634f)))));
+  }
+}
+void net(float* input0, float* output0) {
+r_16_32(buf_0, input0, buf_1, buf_2);
+r_32_16(buf_3, buf_0, buf_4);
+r_4_32(output0, buf_3, buf_5);
+}
+void initialize(float *weights) {
+memcpy(buf_1, weights + 0, 8192);
+memcpy(buf_2, weights + 512, 256);
+memcpy(buf_4, weights + 528, 8192);
+memcpy(buf_5, weights + 1040, 2048);
+}
+int main(int argc, char *argv[]) {
+    // read in the weights from disk
+    FILE *f = fopen("/tmp/tf_weights", "rb");
+    float *weights = (float *)malloc(4672);
+    fread(weights, 1, 4672, f);
+    fclose(f);
+
+    // init the net
+    initialize(weights);
+
+    // test run
+    float input[32];
+    float outputs[4];
+    for (int i = 0; i < 32; i++) scanf("%f", &input[i]);
+    net(input, outputs);
+    printf("%f %f %f %f\n", outputs[0], outputs[1], outputs[2], outputs[3]);
+  }

docs/_source/_static/_output/tinygrad/efficientnet.py.txt

@@ -0,0 +1,2 @@
+281 8.961814 tabby, tabby cat
+did inference in 1105.86 ms

docs/_source/_static/_output/tinygrad/f16_w_uint32.py.txt

@@ -0,0 +1,5 @@
+[0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
+ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
+ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
+ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.
+ 0. 0. 0. 0.]

docs/_source/_static/_output/tinygrad/handcode_resnet50_opt.py.txt

@@ -0,0 +1,61 @@
+optimizing for GPU
+***    2.00 ms : kernel  0 r_64_8_7_7_2_16_4_3_7_4_4_7           [49, 8, 64]        [4, 16, 2]   takes    2.00 ms,   7772 GFLOPS
+***    2.31 ms : kernel  1 r_2048_7_7_2_8_8_3_3                  [7, 7, 2048]       [8, 8, 2]    takes    0.31 ms,    370 GFLOPS
+***    2.49 ms : kernel  2 r_64_2_49_8_16_16_4_4_4               [49, 2, 64]        [16, 8]      takes    0.18 ms,   9548 GFLOPS
+***    3.46 ms : kernel  3 r_64_2_7_7_8_8_2_64_4_4_3_3           [49, 2, 64]        [2, 8, 8]    takes    0.96 ms,  15461 GFLOPS
+***    4.09 ms : kernel  4 r_64_8_49_8_16_16_4_4_4               [49, 8, 64]        [16, 8]      takes    0.64 ms,  10300 GFLOPS
+***    4.79 ms : kernel  5 r_64_8_49_8_16_16_4_4_4n1             [49, 8, 64]        [16, 8]      takes    0.69 ms,  10432 GFLOPS
+***    5.40 ms : kernel  6 r_64_2_49_8_16_64_4_4_4               [49, 2, 64]        [16, 8]      takes    0.61 ms,  10976 GFLOPS
+***    6.36 ms : kernel  7 r_64_2_7_7_8_8_2_64_4_4_3_3n1         [49, 2, 64]        [2, 8, 8]    takes    0.96 ms,  15461 GFLOPS
+***    7.04 ms : kernel  8 r_64_8_49_8_16_16_4_4_4n2             [49, 8, 64]        [16, 8]      takes    0.68 ms,  10321 GFLOPS
+***    7.65 ms : kernel  9 r_64_2_49_8_16_64_4_4_4n1             [49, 2, 64]        [16, 8]      takes    0.61 ms,  10976 GFLOPS
+***    8.62 ms : kernel 10 r_64_2_7_7_8_8_2_64_4_4_3_3n2         [49, 2, 64]        [2, 8, 8]    takes    0.96 ms,  15461 GFLOPS
+***    9.30 ms : kernel 11 r_64_8_49_8_16_16_4_4_4n3             [49, 8, 64]        [16, 8]      takes    0.68 ms,  10321 GFLOPS
+***   10.15 ms : kernel 12 r_64_4_49_8_16_64_4_4_4               [49, 4, 64]        [16, 8]      takes    0.85 ms,  15668 GFLOPS
+***   11.66 ms : kernel 13 r_32_2_7_7_2_16_4_128_4_4_3_3         [49, 2, 32]        [4, 16, 2]   takes    1.51 ms,   9851 GFLOPS
+***   13.29 ms : kernel 14 r_32_8_7_7_2_16_4_64_4_4_4            [49, 8, 32]        [4, 16, 2]   takes    1.63 ms,   8048 GFLOPS
+***   14.12 ms : kernel 15 r_32_8_49_2_16_4_32_4_4_4             [49, 8, 32]        [4, 16, 2]   takes    0.83 ms,   8304 GFLOPS
+***   15.00 ms : kernel 16 r_32_2_49_2_16_4_128_4_4_4            [49, 2, 32]        [4, 16, 2]   takes    0.88 ms,   7567 GFLOPS
+***   16.28 ms : kernel 17 r_32_2_7_7_2_16_4_128_4_4_3_3n1       [49, 2, 32]        [4, 16, 2]   takes    1.28 ms,  11633 GFLOPS
+***   17.11 ms : kernel 18 r_32_8_49_2_16_4_32_4_4_4n1           [49, 8, 32]        [4, 16, 2]   takes    0.83 ms,   8156 GFLOPS
+***   17.99 ms : kernel 19 r_32_2_49_2_16_4_128_4_4_4n1          [49, 2, 32]        [4, 16, 2]   takes    0.88 ms,   7567 GFLOPS
+***   19.26 ms : kernel 20 r_32_2_7_7_2_16_4_128_4_4_3_3n2       [49, 2, 32]        [4, 16, 2]   takes    1.28 ms,  11633 GFLOPS
+***   20.10 ms : kernel 21 r_32_8_49_2_16_4_32_4_4_4n2           [49, 8, 32]        [4, 16, 2]   takes    0.83 ms,   8156 GFLOPS
+***   20.97 ms : kernel 22 r_32_2_49_2_16_4_128_4_4_4n2          [49, 2, 32]        [4, 16, 2]   takes    0.88 ms,   7567 GFLOPS
+***   22.25 ms : kernel 23 r_32_2_7_7_2_16_4_128_4_4_3_3n3       [49, 2, 32]        [4, 16, 2]   takes    1.28 ms,  11633 GFLOPS
+***   23.09 ms : kernel 24 r_32_8_49_2_16_4_32_4_4_4n3           [49, 8, 32]        [4, 16, 2]   takes    0.83 ms,   8156 GFLOPS
+***   24.42 ms : kernel 25 r_32_4_49_2_16_4_128_4_4_4            [49, 4, 32]        [4, 16, 2]   takes    1.33 ms,   9942 GFLOPS
+***   26.54 ms : kernel 26 r_16_4_7_7_16_2_2_256_4_4_3_3         [49, 4, 16]        [2, 2, 16]   takes    2.12 ms,   6986 GFLOPS
+***   29.37 ms : kernel 27 r_16_16_7_7_16_2_2_128_4_4_4          [49, 16, 16]       [2, 2, 16]   takes    2.83 ms,   4642 GFLOPS
+***   30.70 ms : kernel 28 r_8_16_49_8_16_64_4_4_4               [49, 16, 8]        [16, 8]      takes    1.32 ms,   5090 GFLOPS
+***   32.19 ms : kernel 29 r_8_4_49_8_16_256_4_4_4               [49, 4, 8]         [16, 8]      takes    1.49 ms,   4426 GFLOPS
+***   33.73 ms : kernel 30 r_16_4_7_7_16_2_2_256_4_4_3_3n1       [49, 4, 16]        [2, 2, 16]   takes    1.54 ms,   9654 GFLOPS
+***   35.10 ms : kernel 31 r_8_16_49_8_16_64_4_4_4n1             [49, 16, 8]        [16, 8]      takes    1.38 ms,   4860 GFLOPS
+***   36.59 ms : kernel 32 r_8_4_49_8_16_256_4_4_4n1             [49, 4, 8]         [16, 8]      takes    1.49 ms,   4426 GFLOPS
+***   38.13 ms : kernel 33 r_16_4_7_7_16_2_2_256_4_4_3_3n2       [49, 4, 16]        [2, 2, 16]   takes    1.54 ms,   9654 GFLOPS
+***   39.51 ms : kernel 34 r_8_16_49_8_16_64_4_4_4n2             [49, 16, 8]        [16, 8]      takes    1.38 ms,   4860 GFLOPS
+***   41.00 ms : kernel 35 r_8_4_49_8_16_256_4_4_4n2             [49, 4, 8]         [16, 8]      takes    1.49 ms,   4426 GFLOPS
+***   42.53 ms : kernel 36 r_16_4_7_7_16_2_2_256_4_4_3_3n3       [49, 4, 16]        [2, 2, 16]   takes    1.54 ms,   9654 GFLOPS
+***   43.91 ms : kernel 37 r_8_16_49_8_16_64_4_4_4n3             [49, 16, 8]        [16, 8]      takes    1.38 ms,   4860 GFLOPS
+***   45.40 ms : kernel 38 r_8_4_49_8_16_256_4_4_4n3             [49, 4, 8]         [16, 8]      takes    1.49 ms,   4426 GFLOPS
+***   46.94 ms : kernel 39 r_16_4_7_7_16_2_2_256_4_4_3_3n4       [49, 4, 16]        [2, 2, 16]   takes    1.54 ms,   9654 GFLOPS
+***   48.32 ms : kernel 40 r_8_16_49_8_16_64_4_4_4n4             [49, 16, 8]        [16, 8]      takes    1.38 ms,   4860 GFLOPS
+***   49.81 ms : kernel 41 r_8_4_49_8_16_256_4_4_4n4             [49, 4, 8]         [16, 8]      takes    1.49 ms,   4426 GFLOPS
+***   51.34 ms : kernel 42 r_16_4_7_7_16_2_2_256_4_4_3_3n5       [49, 4, 16]        [2, 2, 16]   takes    1.54 ms,   9654 GFLOPS
+***   52.72 ms : kernel 43 r_8_16_49_8_16_64_4_4_4n5             [49, 16, 8]        [16, 8]      takes    1.38 ms,   4860 GFLOPS
+***   55.75 ms : kernel 44 r_8_8_49_8_16_256_4_4_4               [49, 8, 8]         [16, 8]      takes    3.03 ms,   4363 GFLOPS
+***   57.27 ms : kernel 45 r_8_8_8_16_512_3_3_7_7_4              [8, 8]             [16, 8]      takes    1.52 ms,   9721 GFLOPS
+***   61.02 ms : kernel 46 r_2_32_7_7_8_16_256_4_4_4             [49, 32, 2]        [16, 8]      takes    3.75 ms,   3506 GFLOPS
+***   62.81 ms : kernel 47 r_2_32_49_8_16_128_4_4_4              [49, 32, 2]        [16, 8]      takes    1.78 ms,   3732 GFLOPS
+***   64.99 ms : kernel 48 r_2_8_49_8_16_512_4_4_4               [49, 8, 2]         [16, 8]      takes    2.18 ms,   3019 GFLOPS
+***   66.95 ms : kernel 49 r_8_8_8_16_512_3_3_7_7_4n1            [8, 8]             [16, 8]      takes    1.96 ms,   7570 GFLOPS
+***   68.78 ms : kernel 50 r_2_32_49_8_16_128_4_4_4n1            [49, 32, 2]        [16, 8]      takes    1.83 ms,   3619 GFLOPS
+***   70.96 ms : kernel 51 r_2_8_49_8_16_512_4_4_4n1             [49, 8, 2]         [16, 8]      takes    2.18 ms,   3019 GFLOPS
+***   72.92 ms : kernel 52 r_8_8_8_16_512_3_3_7_7_4n2            [8, 8]             [16, 8]      takes    1.96 ms,   7570 GFLOPS
+***   74.75 ms : kernel 53 r_2_32_49_8_16_128_4_4_4n2            [49, 32, 2]        [16, 8]      takes    1.83 ms,   3619 GFLOPS
+***   74.97 ms : kernel 54 r_1024_32_49_4                        [1024]             [32]         takes    0.22 ms,     30 GFLOPS
+***   75.15 ms : kernel 55 r_125_16_2_512_4_4_4                  [125]              [2, 16]      takes    0.17 ms,   1503 GFLOPS
+***   75.16 ms : kernel 56 r_2_32_250_4                          [2]                [32]         takes    0.01 ms,      5 GFLOPS
+***   75.20 ms : kernel 57 r_2_32_250_4n1                        [2]                [32]         takes    0.04 ms,      7 GFLOPS
+***   75.20 ms : kernel 58 E_2_125_32_2_4                        [125, 2]           [2, 32]      takes    0.00 ms,     42 GFLOPS
+******* total 75.20 ms,   7037 GFLOPS

docs/_source/_static/_output/tinygrad/llama.py.txt

@@ -0,0 +1,18 @@
+using GPU backend
+using LLaMA-7B model
+Traceback (most recent call last):
+  File "/home/jebba/devel/tinygrad/tinygrad/examples/llama.py", line 386, in <module>
+    llama = LLaMa.build(MODEL_PATH, TOKENIZER_PATH, model_gen=args.gen, model_size=args.size, quantize=args.quantize, device=device)
+            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/jebba/devel/tinygrad/tinygrad/examples/llama.py", line 155, in build
+    sp_model = SentencePieceProcessor(model_file=str(tokenizer_path))
+               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/jebba/devel/tinygrad/tinygrad/venv/lib/python3.11/site-packages/sentencepiece/__init__.py", line 447, in Init
+    self.Load(model_file=model_file, model_proto=model_proto)
+  File "/home/jebba/devel/tinygrad/tinygrad/venv/lib/python3.11/site-packages/sentencepiece/__init__.py", line 905, in Load
+    return self.LoadFromFile(model_file)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/jebba/devel/tinygrad/tinygrad/venv/lib/python3.11/site-packages/sentencepiece/__init__.py", line 310, in LoadFromFile
+    return _sentencepiece.SentencePieceProcessor_LoadFromFile(self, arg)
+           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+OSError: Not found: "/home/jebba/devel/tinygrad/tinygrad/weights/LLaMA/tokenizer.model": No such file or directory Error #2

docs/_source/_static/_output/tinygrad/mixtral.py.txt

@@ -0,0 +1,8 @@
+Traceback (most recent call last):
+  File "/home/jebba/devel/tinygrad/tinygrad/examples/mixtral.py", line 33, in <module>
+    state = torch_load(args.weights + "/consolidated.00.pth.b")
+            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/home/jebba/devel/tinygrad/tinygrad/tinygrad/nn/state.py", line 77, in torch_load
+    t = Tensor.empty(os.stat(fn).st_size, dtype=dtypes.uint8, device=f"disk:{fn}")
+                     ^^^^^^^^^^^
+FileNotFoundError: [Errno 2] No such file or directory: '/home/jebba/devel/tinygrad/tinygrad/weights/mixtral-8x7b-32kseqlen/consolidated.00.pth.b'

docs/_source/locale/en/LC_MESSAGES/output.po

@@ -8,7 +8,7 @@ msgid ""
 msgstr ""
 "Project-Id-Version: tinyrocs 0\n"
 "Report-Msgid-Bugs-To: \n"
-"POT-Creation-Date: 2024-02-06 11:06-0700\n"
+"POT-Creation-Date: 2024-02-06 11:14-0700\n"
 "PO-Revision-Date: YEAR-MO-DA HO:MI+ZONE\n"
 "Last-Translator: FULL NAME <EMAIL@ADDRESS>\n"
 "Language: en\n"
@@ -128,3 +128,51 @@ msgstr ""
 #: ../../../_source/output.rst:108
 msgid "``python examples/coder.py``"
 msgstr ""
+
+#: ../../../_source/output.rst:114
+msgid "``python examples/compile_efficientnet.py``"
+msgstr ""
+
+#: ../../../_source/output.rst:120
+msgid "``python examples/compile_tensorflow.py.txt``"
+msgstr ""
+
+#: ../../../_source/output.rst:126
+msgid "``python examples/conversation.py``"
+msgstr ""
+
+#: ../../../_source/output.rst:132
+msgid "``python examples/efficientnet.py``"
+msgstr ""
+
+#: ../../../_source/output.rst:138
+msgid "``python examples/f16_w_uint32.py``"
+msgstr ""
+
+#: ../../../_source/output.rst:144
+msgid "``python examples/gpt2.py``"
+msgstr ""
+
+#: ../../../_source/output.rst:150
+msgid "``python examples/handcode_resnet50_opt.py``"
+msgstr ""
+
+#: ../../../_source/output.rst:156
+msgid "``python examples/hlb_cifar10.py``"
+msgstr ""
+
+#: ../../../_source/output.rst:162
+msgid "``python examples/llama.py``"
+msgstr ""
+
+#: ../../../_source/output.rst:168
+msgid "``python examples/mask_rcnn.py``"
+msgstr ""
+
+#: ../../../_source/output.rst:174
+msgid "``python examples/mixtral.py``"
+msgstr ""
+
+#: ../../../_source/output.rst:180
+msgid "``python examples/mnist_gan.py``"
+msgstr ""

docs/_source/output.rst

@@ -110,3 +110,75 @@ Note, while these examples were running, builds were also running, hitting
 .. literalinclude:: _static/_output/tinygrad/coder.py.txt
   :language: output
 
+``python examples/compile_efficientnet.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/compile_efficientnet.py.txt
+  :language: output
+
+``python examples/compile_tensorflow.py.txt``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/compile_tensorflow.py.txt
+  :language: output
+
+``python examples/conversation.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/conversation.py.txt
+  :language: output
+
+``python examples/efficientnet.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/efficientnet.py.txt
+  :language: output
+
+``python examples/f16_w_uint32.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/f16_w_uint32.py.txt
+  :language: output
+
+``python examples/gpt2.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/gpt2.py.txt
+  :language: output
+
+``python examples/handcode_resnet50_opt.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/handcode_resnet50_opt.py.txt
+  :language: output
+
+``python examples/hlb_cifar10.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/hlb_cifar10.py.txt
+  :language: output
+
+``python examples/llama.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/llama.py.txt
+  :language: output
+
+``python examples/mask_rcnn.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/mask_rcnn.py.txt
+  :language: output
+
+``python examples/mixtral.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/mixtral.py.txt
+  :language: output
+
+``python examples/mnist_gan.py``
+----------------------------
+
+.. literalinclude:: _static/_output/tinygrad/mnist_gan.py.txt
+  :language: output
+