vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llama : add support for Chameleon (#8543)

Browse source 
* convert chameleon hf to gguf

* add chameleon tokenizer tests

* fix lint

* implement chameleon graph

* add swin norm param

* return qk norm weights and biases to original format

* implement swin norm

* suppress image token output

* rem tabs

* add comment to conversion

* fix ci

* check for k norm separately

* adapt to new lora implementation

* fix layer input for swin norm

* move swin_norm in gguf writer

* add comment regarding special token regex in chameleon pre-tokenizer

* Update src/llama.cpp

Co-authored-by: compilade <git@compilade.net>

* fix punctuation regex in chameleon pre-tokenizer (@compilade)

Co-authored-by: compilade <git@compilade.net>

* fix lint

* trigger ci

---------

Co-authored-by: compilade <git@compilade.net>
This commit is contained in:
nopperl 2024-09-28 12:08:43 +00:00 • committed by GitHub
parent 43bcdd9703
commit 9a913110cf
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
10 changed files with 505 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

14
src/llama-vocab.cpp
View file

@ -450,6 +450,20 @@ struct llm_tokenizer_bpe {
"[^\\r\\n\\p{L}\\p{N}]?((?=[\\p{L}])([^a-z]))*((?=[\\p{L}])([^A-Z]))+|[^\\r\\n\\p{L}\\p{N}]?((?=[\\p{L}])([^a-z]))+((?=[\\p{L}])([^A-Z]))*|\\p{N}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n/]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
};
break;
case LLAMA_VOCAB_PRE_TYPE_CHAMELEON:
// Note: in theory, the special token (sentinel and image token) regex_exprs below
// are unnecessary, as they are split in `tokenizer_st_partition` anyway.
// However, since the upstream pre-tokenizer uses them, they are also
// included here (see https://huggingface.co/facebook/chameleon-7b).
regex_exprs = {
"<sentinel:[0-9]+>", // Sentinel tokens
"(IMGIMG)((A|B|C|D|E|F|G|H|I){1,4})Z", // Image tokens
"([\\t\\n]| | )", // directly from tokenizer.json
"\\p{N}", // Individual digits
"[\\p{P}!-/:-@\\[-`{-~]", // Punctuation, Isolated
"'s|'t|'re|'ve|'m|'ll|'d| ?\\p{L}+| ?\\p{N}+| ?[^\\s\\p{L}\\p{N}]+|\\s+(?!\\S)",
};
break;
default:
// default regex for BPE tokenization pre-processing
regex_exprs = {

263
src/llama.cpp
View file

@ -216,6 +216,7 @@ enum llm_arch {
LLM_ARCH_RWKV6,
LLM_ARCH_GRANITE,
LLM_ARCH_GRANITE_MOE,
LLM_ARCH_CHAMELEON,
LLM_ARCH_UNKNOWN,
};
@ -268,6 +269,7 @@ static const std::map<llm_arch, const char *> LLM_ARCH_NAMES = {
{ LLM_ARCH_RWKV6, "rwkv6" },
{ LLM_ARCH_GRANITE, "granite" },
{ LLM_ARCH_GRANITE_MOE, "granitemoe" },
{ LLM_ARCH_CHAMELEON, "chameleon" },
{ LLM_ARCH_UNKNOWN, "(unknown)" },
};
@ -304,6 +306,7 @@ enum llm_kv {
LLM_KV_DECODER_START_TOKEN_ID,
LLM_KV_ATTN_LOGIT_SOFTCAPPING,
LLM_KV_FINAL_LOGIT_SOFTCAPPING,
LLM_KV_SWIN_NORM,
LLM_KV_RESCALE_EVERY_N_LAYERS,
LLM_KV_TIME_MIX_EXTRA_DIM,
LLM_KV_TIME_DECAY_EXTRA_DIM,
@ -411,6 +414,7 @@ static const std::map<llm_kv, const char *> LLM_KV_NAMES = {
{ LLM_KV_DECODER_START_TOKEN_ID, "%s.decoder_start_token_id" },
{ LLM_KV_ATTN_LOGIT_SOFTCAPPING, "%s.attn_logit_softcapping" },
{ LLM_KV_FINAL_LOGIT_SOFTCAPPING, "%s.final_logit_softcapping" },
{ LLM_KV_SWIN_NORM, "%s.swin_norm" },
{ LLM_KV_RESCALE_EVERY_N_LAYERS, "%s.rescale_every_n_layers" },
{ LLM_KV_TIME_MIX_EXTRA_DIM, "%s.time_mix_extra_dim" },
{ LLM_KV_TIME_DECAY_EXTRA_DIM, "%s.time_decay_extra_dim" },
@ -1499,6 +1503,25 @@ static const std::map<llm_arch, std::map<llm_tensor, std::string>> LLM_TENSOR_NA
{ LLM_TENSOR_FFN_UP_EXPS, "blk.%d.ffn_up_exps" },
},
},
{
LLM_ARCH_CHAMELEON,
{
{ LLM_TENSOR_TOKEN_EMBD, "token_embd" },
{ LLM_TENSOR_OUTPUT_NORM, "output_norm" },
{ LLM_TENSOR_OUTPUT, "output" },
{ LLM_TENSOR_ATTN_NORM, "blk.%d.attn_norm" },
{ LLM_TENSOR_ATTN_Q, "blk.%d.attn_q" },
{ LLM_TENSOR_ATTN_K, "blk.%d.attn_k" },
{ LLM_TENSOR_ATTN_V, "blk.%d.attn_v" },
{ LLM_TENSOR_ATTN_OUT, "blk.%d.attn_output" },
{ LLM_TENSOR_FFN_NORM, "blk.%d.ffn_norm" },
{ LLM_TENSOR_FFN_GATE, "blk.%d.ffn_gate" },
{ LLM_TENSOR_FFN_DOWN, "blk.%d.ffn_down" },
{ LLM_TENSOR_FFN_UP, "blk.%d.ffn_up" },
{ LLM_TENSOR_ATTN_Q_NORM, "blk.%d.attn_q_norm" },
{ LLM_TENSOR_ATTN_K_NORM, "blk.%d.attn_k_norm" },
},
},
{
LLM_ARCH_UNKNOWN,
{
@ -2362,6 +2385,7 @@ struct llama_hparams {
bool vocab_only;
bool rope_finetuned;
bool use_par_res;
bool swin_norm;
uint32_t n_vocab;
uint32_t n_ctx_train; // context size the model was trained on
@ -6084,6 +6108,18 @@ static void llm_load_hparams(
default: model.type = e_model::MODEL_UNKNOWN;
}
} break;
case LLM_ARCH_CHAMELEON:
{
ml.get_key(LLM_KV_ATTENTION_LAYERNORM_RMS_EPS, hparams.f_norm_rms_eps);
hparams.f_norm_eps = 1e-5; // eps for qk-norm, torch default
ml.get_key(LLM_KV_SWIN_NORM, hparams.swin_norm);
switch (hparams.n_layer) {
case 32: model.type = e_model::MODEL_7B; break;
case 48: model.type = e_model::MODEL_34B; break;
default: model.type = e_model::MODEL_UNKNOWN;
}
} break;
default: (void)0;
}
@ -6341,6 +6377,11 @@ static void llm_load_vocab(
} else if (
tokenizer_pre == "exaone") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_EXAONE;
} else if (
tokenizer_pre == "chameleon") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_CHAMELEON;
vocab.tokenizer_add_bos = true;
vocab.tokenizer_clean_spaces = false;
} else {
throw std::runtime_error(format("unknown pre-tokenizer type: '%s'", tokenizer_pre.c_str()));
}
@ -8728,6 +8769,45 @@ static bool llm_load_tensors(
}
} break;
case LLM_ARCH_CHAMELEON:
{
model.tok_embd = ml.create_tensor(ctx_input, tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab});
// output
{
model.output_norm = ml.create_tensor(ctx_output, tn(LLM_TENSOR_OUTPUT_NORM, "weight"), {n_embd});
model.output = ml.create_tensor(ctx_output_split, tn(LLM_TENSOR_OUTPUT, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_NOT_REQUIRED);
// if output is NULL, init from the input tok embed
if (model.output == NULL) {
model.output = ml.create_tensor(ctx_output, tn(LLM_TENSOR_TOKEN_EMBD, "weight"), {n_embd, n_vocab}, llama_model_loader::TENSOR_DUPLICATED);
}
}
for (int i = 0; i < n_layer; ++i) {
ggml_context * ctx_layer = ctx_for_layer(i);
ggml_context * ctx_split = ctx_for_layer_split(i);
auto & layer = model.layers[i];
layer.attn_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_NORM, "weight", i), {n_embd});
layer.attn_q_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_Q_NORM, "weight", i), {n_embd_head_k, n_head});
layer.attn_k_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_K_NORM, "weight", i), {n_embd_head_k, n_head_kv});
layer.attn_q_norm_b = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_Q_NORM, "bias", i), {n_embd_head_k, n_head}, llama_model_loader::TENSOR_NOT_REQUIRED);
layer.attn_k_norm_b = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_ATTN_K_NORM, "bias", i), {n_embd_head_k, n_head_kv}, llama_model_loader::TENSOR_NOT_REQUIRED);
layer.wq = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_Q, "weight", i), {n_embd, n_embd});
layer.wk = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_K, "weight", i), {n_embd, n_embd_gqa});
layer.wv = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_V, "weight", i), {n_embd, n_embd_gqa});
layer.wo = ml.create_tensor(ctx_split, tn(LLM_TENSOR_ATTN_OUT, "weight", i), {n_embd, n_embd});
layer.ffn_norm = ml.create_tensor(ctx_layer, tn(LLM_TENSOR_FFN_NORM, "weight", i), {n_embd});
layer.ffn_gate = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_GATE, "weight", i), {n_embd, n_ff});
layer.ffn_down = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_DOWN, "weight", i), { n_ff, n_embd});
layer.ffn_up = ml.create_tensor(ctx_split, tn(LLM_TENSOR_FFN_UP, "weight", i), {n_embd, n_ff});
}
} break;
default:
throw std::runtime_error("unknown architecture");
}
@ -15872,6 +15952,184 @@ struct llm_build_context {
return gf;
}
// ref: https://github.com/facebookresearch/chameleon
// based on the original build_llama() function, changes:
// * qk-norm
// * swin-norm
// * removed bias
// * removed MoE
struct ggml_cgraph * build_chameleon() {
struct ggml_cgraph * gf = ggml_new_graph_custom(ctx0, llama_model_max_nodes(model), false);
// mutable variable, needed during the last layer of the computation to skip unused tokens
int32_t n_tokens = this->n_tokens;
const int64_t n_embd_head = hparams.n_embd_head_v;
GGML_ASSERT(n_embd_head == hparams.n_embd_head_k);
GGML_ASSERT(n_embd_head == hparams.n_rot);
struct ggml_tensor * cur;
struct ggml_tensor * inpL;
inpL = llm_build_inp_embd(ctx0, lctx, hparams, batch, model.tok_embd, cb);
// inp_pos - contains the positions
struct ggml_tensor * inp_pos = build_inp_pos();
// KQ_mask (mask for 1 head, it will be broadcasted to all heads)
struct ggml_tensor * KQ_mask = build_inp_KQ_mask();
for (int il = 0; il < n_layer; ++il) {
struct ggml_tensor * inpSA = inpL;
// norm
if (hparams.swin_norm) {
cur = inpL;
} else {
cur = llm_build_norm(ctx0, inpL, hparams,
model.layers[il].attn_norm, NULL,
LLM_NORM_RMS, cb, il);
cb(cur, "attn_norm", il);
}
// self-attention
{
// compute Q and K and RoPE them
struct ggml_tensor * Qcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wq, cur);
cb(Qcur, "Qcur", il);
struct ggml_tensor * Kcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wk, cur);
cb(Kcur, "Kcur", il);
struct ggml_tensor * Vcur = llm_build_lora_mm(lctx, ctx0, model.layers[il].wv, cur);
cb(Vcur, "Vcur", il);
if (model.layers[il].attn_q_norm) {
Qcur = ggml_view_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens,
ggml_element_size(Qcur) * n_embd_head,
ggml_element_size(Qcur) * n_embd_head * n_head,
0);
cb(Qcur, "Qcur", il);
Qcur = llm_build_norm(ctx0, Qcur, hparams,
model.layers[il].attn_q_norm,
model.layers[il].attn_q_norm_b,
LLM_NORM, cb, il);
cb(Qcur, "Qcur", il);
}
if (model.layers[il].attn_k_norm) {
Kcur = ggml_view_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens,
ggml_element_size(Kcur) * n_embd_head,
ggml_element_size(Kcur) * n_embd_head * n_head_kv,
0);
cb(Kcur, "Kcur", il);
Kcur = llm_build_norm(ctx0, Kcur, hparams,
model.layers[il].attn_k_norm,
model.layers[il].attn_k_norm_b,
LLM_NORM, cb, il);
cb(Kcur, "Kcur", il);
}
Qcur = ggml_rope_ext(
ctx0, ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens), inp_pos, nullptr,
n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
ext_factor, attn_factor, beta_fast, beta_slow
);
cb(Qcur, "Qcur", il);
Kcur = ggml_rope_ext(
ctx0, ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens), inp_pos, nullptr,
n_rot, rope_type, n_ctx_orig, freq_base, freq_scale,
ext_factor, attn_factor, beta_fast, beta_slow
);
cb(Kcur, "Kcur", il);
cur = llm_build_kv(ctx0, lctx, kv_self, gf,
model.layers[il].wo, nullptr,
Kcur, Vcur, Qcur, KQ_mask, n_tokens, kv_head, n_kv, 1.0f/sqrtf(float(n_embd_head)), cb, il);
if (hparams.swin_norm) {
cur = llm_build_norm(ctx0, cur, hparams,
model.layers[il].attn_norm, NULL,
LLM_NORM_RMS, cb, il);
}
}
if (il == n_layer - 1) {
// skip computing output for unused tokens
struct ggml_tensor * inp_out_ids = build_inp_out_ids();
n_tokens = n_outputs;
cur = ggml_get_rows(ctx0, cur, inp_out_ids);
inpSA = ggml_get_rows(ctx0, inpSA, inp_out_ids);
}
struct ggml_tensor * ffn_inp = ggml_add(ctx0, cur, inpSA);
cb(ffn_inp, "ffn_inp", il);
// feed-forward network
if (!hparams.swin_norm) {
cur = llm_build_norm(ctx0, ffn_inp, hparams,
model.layers[il].ffn_norm, NULL,
LLM_NORM_RMS, cb, il);
cb(cur, "ffn_norm", il);
}
cur = llm_build_ffn(ctx0, lctx, cur,
model.layers[il].ffn_up, NULL, NULL,
model.layers[il].ffn_gate, NULL, NULL,
model.layers[il].ffn_down, NULL, NULL,
NULL,
LLM_FFN_SILU, LLM_FFN_PAR, cb, il);
cb(cur, "ffn_out", il);
if (hparams.swin_norm) {
cur = llm_build_norm(ctx0, cur, hparams,
model.layers[il].ffn_norm, NULL,
LLM_NORM_RMS, cb, il);
cb(cur, "ffn_norm", il);
}
cur = ggml_add(ctx0, cur, ffn_inp);
cb(cur, "ffn_out", il);
cur = lctx.cvec.apply_to(ctx0, cur, il);
cb(cur, "l_out", il);
// input for next layer
inpL = cur;
}
cur = inpL;
cur = llm_build_norm(ctx0, cur, hparams,
model.output_norm, NULL,
LLM_NORM_RMS, cb, -1);
cb(cur, "result_norm", -1);
// lm_head
cur = llm_build_lora_mm(lctx, ctx0, model.output, cur);
cb(cur, "result_output_with_img_logits", -1);
// TODO: this suppresses the output of image tokens, which is required to enable text-only outputs.
// Needs to be removed once image outputs are supported.
int img_token_end_idx = 8196;
int img_token_start_idx = 4;
int num_img_tokens = img_token_end_idx - img_token_start_idx;
// creates 1d tensor of size num_img_tokens and values -FLT_MAX,
// which ensures that text token values are always at least larger than image token values
struct ggml_tensor * img_logits = ggml_new_tensor_1d(ctx0, GGML_TYPE_F32, num_img_tokens);
img_logits = ggml_clamp(ctx0, img_logits, -FLT_MAX, -FLT_MAX);
cb(img_logits, "img_logits", -1);
cur = ggml_set_1d(ctx0, cur, img_logits, ggml_element_size(cur) * img_token_start_idx);
cb(cur, "result_output", -1);
ggml_build_forward_expand(gf, cur);
return gf;
}
};
static struct ggml_cgraph * llama_build_graph_defrag(llama_context & lctx, const std::vector<uint32_t> & ids) {
@ -16132,6 +16390,10 @@ static struct ggml_cgraph * llama_build_graph(
{
result = llm.build_rwkv6();
} break;
case LLM_ARCH_CHAMELEON:
{
result = llm.build_chameleon();
} break;
default:
GGML_ABORT("fatal error");
}
@ -19257,6 +19519,7 @@ enum llama_rope_type llama_rope_type(const struct llama_model * model) {
case LLM_ARCH_CHATGLM:
case LLM_ARCH_GRANITE:
case LLM_ARCH_GRANITE_MOE:
case LLM_ARCH_CHAMELEON:
return LLAMA_ROPE_TYPE_NORM;
// the pairs of head values are offset by n_rot/2