Run frontier MoE models on hardware you already own — pure C, zero deps, experts streamed from disk. Tiny engine, immense model. 🐦
C
qwen-image engine: bounded by what a hostile model file or request can ask
JustVugg126c5f5fb07f
diff --git a/c/qwenimage.c b/c/qwenimage.cindex 1a44a735..4f025a64 100644--- a/c/qwenimage.c+++ b/c/qwenimage.c@@ -260,8 +260,13 @@ static void te_config(Te *te, const char *model){te->theta = (float)jnum(rp, "rope_theta", 5000000.0);}json_free(root); free(arena);- if (te->hidden <= 0 || te->layers <= 0 || te->heads <= 0 || te->kv <= 0 || te->hd <= 0 ||- te->heads % te->kv || te->inter <= 0 || te->vocab <= 0 || te->hd % 2) {+ /* Model files come from mirrors nobody vouches for: every dimension that+ * sizes a buffer is bounded, and products are formed in 64 bits, so a config+ * cannot wrap an int into a small allocation that a loop then overruns. */+ if (te->hidden <= 0 || te->hidden > 65536 || te->layers <= 0 || te->layers > 1024 ||+ te->heads <= 0 || te->heads > 1024 || te->kv <= 0 || te->hd <= 0 || te->hd > 1024 ||+ te->heads % te->kv || te->inter <= 0 || te->inter > (1 << 18) || te->vocab <= 0 ||+ te->vocab > (1 << 22) || te->hd % 2 || (int64_t)te->heads * te->hd > 65536) {fprintf(stderr, "[qwenimage] text_encoder/config.json: unsupported shape\n"); exit(1);}}@@ -438,8 +443,10 @@ static void dit_config(Dit *d, const char *model){d->layers = (int)jnum(c, "num_layers", 0);d->heads = (int)jnum(c, "num_attention_heads", 0);d->hd = (int)jnum(c, "attention_head_dim", 0);- d->dim = d->heads * d->hd;- d->mlp = d->dim * (int)jnum(c, "mlp_ratio", 3);+ int mlp_ratio = (int)jnum(c, "mlp_ratio", 3);+ int64_t dim64 = (int64_t)d->heads * d->hd;+ d->dim = dim64 > 0 && dim64 <= 65536 ? (int)dim64 : 0;+ d->mlp = mlp_ratio >= 1 && mlp_ratio <= 16 ? d->dim * mlp_ratio : 0;d->in_ch = (int)jnum(c, "in_channels", 64);d->ctx = (int)jnum(c, "context_in_dim", 0);d->eps = (float)jnum(c, "eps", 1e-6);@@ -447,10 +454,15 @@ static void dit_config(Dit *d, const char *model){int patch = (int)jnum(c, "patch_size", 1), outc = (int)jnum(c, "out_channels", d->in_ch);jval *ax = json_get(c, "axes_dims_rope");int ok = ax && ax->t == J_ARR && ax->len == 3;- for (int i = 0; ok && i < 3; i++) d->axes[i] = (int)ax->kids[i]->num;+ for (int i = 0; ok && i < 3; i++) {+ d->axes[i] = ax->kids[i]->t == J_NUM ? (int)ax->kids[i]->num : -1;+ ok = d->axes[i] > 0 && d->axes[i] % 2 == 0; /* each axis fills whole complex pairs */+ }json_free(c); free(arena);- if (!ok || d->axes[0] + d->axes[1] + d->axes[2] != d->hd || d->layers <= 0 || d->heads <= 0 ||- d->hd % 2 || patch != 1 || outc != d->in_ch || d->ctx <= 0 || !d->causal_condition) {+ if (!ok || d->axes[0] + d->axes[1] + d->axes[2] != d->hd || d->layers <= 0 || d->layers > 1024 ||+ d->heads <= 0 || d->heads > 1024 || d->hd <= 0 || d->hd > 1024 || d->hd % 2 || d->dim <= 0 ||+ d->mlp <= 0 || d->in_ch <= 0 || d->in_ch > 1024 || patch != 1 || outc != d->in_ch ||+ d->ctx <= 0 || d->ctx > 65536 || !d->causal_condition) {fprintf(stderr, "[qwenimage] transformer/config.json: unsupported shape ""(needs patch_size 1, causal_condition, axes summing to the head dim)\n");exit(1);@@ -822,6 +834,9 @@ typedef struct { const char *id; int (*cancelled)(const char *id); int preview;/* The picture as it stands after a step: the flow-matching estimate of the clean* latent (x0 = x_t - sigma * v) through a linear latent -> RGB map, one pixel per* latent pixel. 64x3 multiply-adds per token: free next to a DiT step. */+static void json_escape_out(const char *s);+static void send_error(const char *id, const char *m);+static void emit_preview(const Progress *pg, int step, const float *lat, const float *v, float sigma,int gh, int gw, int C){if (!pg || !pg->id || !pg->preview || C != 64) return;@@ -836,8 +851,8 @@ static void emit_preview(const Progress *pg, int step, const float *lat, const frgb[t * 3 + c] = a < 0 ? 0 : a > 255 ? 255 : (uint8_t)a;}}- printf("PREVIEW {\"id\":\"%s\",\"step\":%d,\"width\":%d,\"height\":%d,\"channels\":3,\"bytes\":%zu}\n",- pg->id, step, gw, gh, n);+ printf("PREVIEW {\"id\":\""); json_escape_out(pg->id);+ printf("\",\"step\":%d,\"width\":%d,\"height\":%d,\"channels\":3,\"bytes\":%zu}\n", step, gw, gh, n);fwrite(rgb, 1, n, stdout);putchar('\n');fflush(stdout);@@ -846,8 +861,8 @@ static void emit_preview(const Progress *pg, int step, const float *lat, const fstatic void emit_progress(const Progress *pg, const char *stage, int step, int steps, double t0){if (!pg || !pg->id) return;- printf("PROGRESS {\"id\":\"%s\",\"stage\":\"%s\",\"step\":%d,\"steps\":%d,\"elapsed\":%.2f}\n",- pg->id, stage, step, steps, now_s() - t0);+ printf("PROGRESS {\"id\":\""); json_escape_out(pg->id);+ printf("\",\"stage\":\"%s\",\"step\":%d,\"steps\":%d,\"elapsed\":%.2f}\n", stage, step, steps, now_s() - t0);fflush(stdout);}@@ -862,10 +877,49 @@ static void engine_init(Engine *e, const char *model){fprintf(stderr, "[qwenimage] text encoder hidden %d != transformer context %d\n", e->te.hidden, e->dit.ctx);exit(1);}+ /* The VAE is loaded only when the first image is decoded; its geometry is+ * checked now, so a VAE that does not fit the DiT (latent channels, 16x+ * upsampling, which the image buffers are sized from) fails at start-up+ * and not after a whole denoise. */+ char dir[2100]; snprintf(dir, sizeof dir, "%s/vae", model);+ char *arena = NULL;+ jval *v = read_json(dir, "config.json", &arena);+ jval *dm = json_get(v, "dim_mult");+ int zdim = (int)jnum(v, "z_dim", 0), ok = dm && dm->t == J_ARR && dm->len >= 1 && dm->len <= 8;+ for (int i = 0; ok && i < dm->len; i++) ok = dm->kids[i]->t == J_NUM && dm->kids[i]->num >= 1;+ int scale = ok ? 1 << (dm->len - 1) : 0;+ json_free(v); free(arena);+ if (!ok || scale != 16 || zdim != e->dit.in_ch) {+ fprintf(stderr, "[qwenimage] vae/config.json: z_dim %d and a %dx upsampling do not fit the transformer "+ "(%d latent channels, 16x)\n", zdim, scale, e->dit.in_ch);+ exit(1);+ }+}++#ifdef QI_HAVE_VAE+/* qiv_load plus the same geometry check, against the weights actually loaded. */+static QiVae *engine_vae(Engine *e, char *msg, size_t msgn){+ char dir[2100]; snprintf(dir, sizeof dir, "%s/vae", e->model);+ QiVae *v = qiv_load(dir);+ if (!v) { snprintf(msg, msgn, "cannot load the VAE"); return NULL; }+ if (v->scale != 16 || v->z_dim != e->dit.in_ch) {+ snprintf(msg, msgn, "the VAE (z_dim %d, %dx) does not fit the transformer (%d channels, 16x)",+ v->z_dim, v->scale, e->dit.in_ch);+ qiv_free(v);+ return NULL;+ }+ return v;}+#endif/* The prompt's prefix, cached for the next request with the same prompt. */-static int engine_prefix(Engine *e, const char *prompt, float **emb_out, int *L_out){+/* Prompt tokens, template included. Attention over the prompt keeps a full+ * score matrix, so an unbounded prompt is an unbounded allocation (40000 digits,+ * one token each, took 6.4 GB on the tiny model); the pipeline's own prompts+ * are a few dozen tokens. */+#define QI_MAX_PROMPT_TOKENS 1024++static int engine_prefix(Engine *e, const char *prompt, float **emb_out, int *L_out, char *msg, size_t msgn){if (e->last_prompt && !strcmp(e->last_prompt, prompt) && e->prefix.K && !emb_out) {*L_out = e->prefix.L; return 1;}@@ -873,7 +927,11 @@ static int engine_prefix(Engine *e, const char *prompt, float **emb_out, int *L_free(e->last_prompt); e->last_prompt = NULL;int n, drop;int *ids = te_encode_prompt(&e->te, prompt, &n, &drop);- if (n <= drop) { fprintf(stderr, "[qwenimage] empty prompt after the template\n"); free(ids); return -1; }+ if (n <= drop) { snprintf(msg, msgn, "empty prompt"); free(ids); return -1; }+ if (n > QI_MAX_PROMPT_TOKENS) {+ snprintf(msg, msgn, "prompt too long: %d tokens, the limit is %d", n, QI_MAX_PROMPT_TOKENS);+ free(ids); return -1;+ }te_load(&e->te, e->model);float *h = te_forward(&e->te, ids, n);free(ids);@@ -897,7 +955,7 @@ static int engine_generate(Engine *e, const char *prompt, int width, int height,int gh = height / 16, gw = width / 16, N = gh * gw, C = e->dit.in_ch;emit_progress(pg, "encode", 0, steps, t0);int L;- if (engine_prefix(e, prompt, NULL, &L) < 0) { snprintf(msg, msgn, "empty prompt"); return -1; }+ if (engine_prefix(e, prompt, NULL, &L, msg, msgn) < 0) return -1;double t1 = now_s();float *lat = fmalloc((size_t)N * C), *np = fmalloc((size_t)N * C);randn(lat, (size_t)N * C, seed);@@ -924,11 +982,7 @@ static int engine_generate(Engine *e, const char *prompt, int width, int height,if (rc == 0) {emit_progress(pg, "decode", steps, steps, t0);#ifdef QI_HAVE_VAE- if (!e->vae) {- char dir[2100]; snprintf(dir, sizeof dir, "%s/vae", e->model);- e->vae = qiv_load(dir);- if (!e->vae) { snprintf(msg, msgn, "cannot load the VAE"); rc = -1; }- }+ if (!e->vae && !(e->vae = engine_vae(e, msg, msgn))) rc = -1;if (rc == 0 && qiv_decode(e->vae, lat, gh, gw, rgba, NULL) != 0) { snprintf(msg, msgn, "VAE decode failed"); rc = -1; }#else/* no VAE compiled in: show the first three latent channels, stretched */@@ -952,7 +1006,10 @@ static int check_size(int w, int h, int steps, char *msg, size_t n){snprintf(msg, n, "width and height must be multiples of 32 between 256 and 2048 (got %dx%d)", w, h);return -1;}- if (steps < 1 || steps > 200) { snprintf(msg, n, "steps must be between 1 and 200 (got %d)", steps); return -1; }+ /* One step is refused, as the reference would give NaN: with shift_terminal+ * the schedule's last sigma is also its first, and the stretch divides by+ * 1 - sigma = 0 (diffusers' stretch_shift_to_terminal does the same). */+ if (steps < 2 || steps > 200) { snprintf(msg, n, "steps must be between 2 and 200 (got %d)", steps); return -1; }return 0;}@@ -960,20 +1017,59 @@ static int check_size(int w, int h, int steps, char *msg, size_t n){/* Our own line reader on fd 0: stdio's buffer would hide a CANCEL that arrived* together with the GEN from coli_serve_stdin_ready(). */-static char g_in[1 << 16]; static size_t g_in_n;+/* The buffer grows up to QI_MAX_LINE (the gateway's own body limit). A longer+ * line is read through to its newline and discarded, and read_line hands back+ * "TOOLONG <id>" with the id found at its start, so the client that sent it+ * gets an ERROR instead of waiting forever for an answer to a request the+ * engine never saw. */+#define QI_MAX_LINE ((size_t)4 << 20)+static char *g_in; static size_t g_in_n, g_in_cap;+static int g_skipping; static char g_skip_id[128];++static void note_line_id(const char *buf, size_t n){+ snprintf(g_skip_id, sizeof g_skip_id, "?");+ const char *k = NULL; /* memmem is not in the Windows CRT */+ for (size_t i = 0; i + 6 <= (n < 4096 ? n : 4096); i++)+ if (!memcmp(buf + i, "\"id\":\"", 6)) { k = buf + i; break; }+ if (!k) return;+ k += 6;+ size_t i = 0;+ while (i + 1 < sizeof g_skip_id && k + i < buf + n && k[i] != '"' && k[i] != '\\' && (unsigned char)k[i] >= 0x20) {+ g_skip_id[i] = k[i]; i++;+ }+ g_skip_id[i] = 0;+}+static char *read_line(int block){for (;;) {- char *nl = memchr(g_in, '\n', g_in_n);+ char *nl = g_in_n ? memchr(g_in, '\n', g_in_n) : NULL;if (nl) {size_t len = (size_t)(nl - g_in);- char *line = xmalloc(len + 1);- memcpy(line, g_in, len); line[len] = 0;+ char *line;+ if (g_skipping) { /* the tail of an overlong line */+ g_skipping = 0;+ line = xmalloc(sizeof g_skip_id + 16);+ snprintf(line, sizeof g_skip_id + 16, "TOOLONG %s", g_skip_id);+ } else {+ line = xmalloc(len + 1);+ memcpy(line, g_in, len); line[len] = 0;+ }memmove(g_in, nl + 1, g_in_n - len - 1); g_in_n -= len + 1;return line;}+ if (g_in_n == QI_MAX_LINE) {+ if (!g_skipping) { note_line_id(g_in, g_in_n); g_skipping = 1; }+ g_in_n = 0;+ }if (!block && !coli_serve_stdin_ready()) return NULL;- if (g_in_n == sizeof g_in) { g_in_n = 0; continue; } /* an absurd line: drop it */- ssize_t r = read(0, g_in + g_in_n, sizeof g_in - g_in_n);+ if (g_in_n == g_in_cap) {+ size_t cap = g_in_cap ? g_in_cap * 2 : (size_t)1 << 16;+ if (cap > QI_MAX_LINE) cap = QI_MAX_LINE;+ char *grown = realloc(g_in, cap);+ if (!grown) { fprintf(stderr, "[qwenimage] out of memory reading stdin\n"); exit(1); }+ g_in = grown; g_in_cap = cap;+ }+ ssize_t r = read(0, g_in + g_in_n, g_in_cap - g_in_n);if (r <= 0) return NULL;g_in_n += (size_t)r;}@@ -981,10 +1077,26 @@ static char *read_line(int block){static int g_cancel_pending;static char g_cancel_id[128];+/* Lines that arrive while an image is being made and are not a CANCEL for it+ * (a GEN from a client that gave up waiting, say) are kept for serve_loop,+ * never dropped: a dropped GEN is a client waiting forever for its answer. */+static char *g_later[64]; static int g_later_n;static int serve_cancelled(const char *id){char *line;while ((line = read_line(0))) {- if (!strncmp(line, "CANCEL ", 7)) {+ if (strncmp(line, "CANCEL ", 7)) {+ if (g_later_n < (int)(sizeof g_later / sizeof g_later[0])) { g_later[g_later_n++] = line; continue; }+ if (!strncmp(line, "GEN ", 4)) { /* the queue is full: answer rather than drop */+ char *arena = NULL; jval *j = json_parse(line + 4, &arena);+ jval *v = j ? json_get(j, "id") : NULL;+ send_error(v && v->t == J_STR ? v->str : "?", "engine busy");+ if (j) json_free(j);+ free(arena);+ }+ free(line);+ continue;+ }+ {char *arena = NULL; jval *j = json_parse(line + 7, &arena);jval *v = j ? json_get(j, "id") : NULL;if (v && v->t == J_STR) { snprintf(g_cancel_id, sizeof g_cancel_id, "%s", v->str); g_cancel_pending = 1; }@@ -1014,7 +1126,12 @@ static int serve_loop(Engine *e, int dw, int dh, int dsteps){"\"min_side\":256,\"max_side\":2048,\"multiple\":32}\n", dw, dh, dsteps);fflush(stdout);char *line;- while ((line = read_line(1))) {+ for (;;) {+ if (g_later_n) { /* what came in during the last image */+ line = g_later[0];+ memmove(g_later, g_later + 1, sizeof g_later[0] * (size_t)--g_later_n);+ } else if (!(line = read_line(1))) break;+ if (!strncmp(line, "TOOLONG ", 8)) { send_error(line + 8, "request line too long"); free(line); continue; }if (strncmp(line, "GEN ", 4)) { free(line); continue; } /* CANCEL with nothing running, noise */char *arena = NULL;jval *j = json_parse(line + 4, &arena);@@ -1177,8 +1294,9 @@ static int run_oracle(Engine *e, const char *refdir){dit_step_free(&s);#ifdef QI_HAVE_VAE{- char dir[2100]; snprintf(dir, sizeof dir, "%s/vae", e->model);- QiVae *vae = qiv_load(dir);+ char vmsg[256];+ QiVae *vae = engine_vae(e, vmsg, sizeof vmsg);+ if (!vae) { fprintf(stderr, "[oracle] %s\n", vmsg); fails++; }if (vae && fin) {uint8_t *rgba = xmalloc((size_t)width * height * 4);float *of = fmalloc((size_t)4 * width * height);diff --git a/c/qwenimage_vae.h b/c/qwenimage_vae.hindex fe4ab22b..2739eac3 100644--- a/c/qwenimage_vae.h+++ b/c/qwenimage_vae.h@@ -279,6 +279,14 @@ static QiVae *qiv_load(const char *vae_dir){for (int i = 0; i < v->z_dim; i++) { v->mean[i] = (float)jmean->kids[i]->num; v->std[i] = (float)jstd->kids[i]->num; }int nm = jm->len, dims[QIV_MAX_BLOCKS + 1], tup[QIV_MAX_BLOCKS];+ /* Every width is a product of two config numbers that size buffers and a+ * divisor in qiv_dup_table: positive and bounded, or refused. */+ for (int i = 0; i < nm; i++)+ if (jm->kids[i]->t != J_NUM || jm->kids[i]->num < 1 || jm->kids[i]->num > 64 || base > 4096) {+ snprintf(err, sizeof err, "dim_mult entries must be whole numbers from 1 to 64 (base %d)", base); goto fail; }+ for (int i = 0; i < nm - 1; i++)+ if (jt->kids[i]->t != J_BOOL) { snprintf(err, sizeof err, "temperal_downsample must hold booleans"); goto fail; }+ if (v->z_dim > 4096) { snprintf(err, sizeof err, "z_dim %d is out of range", v->z_dim); goto fail; }dims[0] = base * (int)jm->kids[nm - 1]->num;for (int i = 0; i < nm; i++) dims[i + 1] = base * (int)jm->kids[nm - 1 - i]->num;for (int i = 0; i < nm - 1; i++) tup[i] = jt->kids[nm - 2 - i]->boolean; /* temperal_upsample = reversed */