hugo/DS4Server
1
0
Fork 0
Code Issues 2 Pull Requests Actions Packages Projects Releases Wiki Activity

Support DeepSeek V4 Flash 0731

Browse Source
This commit is contained in:
Georg Bauer
2026-08-29 20:28:50 +02:00
parent ad855b321e
commit f1c177b754
23 changed files with 10510 additions and 2324 deletions
Download Patch File Download Diff File Expand all files Collapse all files

36
native/metal/ds4.h
View File

@@ -150,6 +150,7 @@ typedef struct {
bool glm_mtp_timing;
bool dspark;
bool dspark_strict;
bool dspark_exact_sampling;
bool dspark_confidence_threshold_set;
bool cuda_tensor_parallel;
bool ssd_streaming;
@@ -158,6 +159,8 @@ typedef struct {
bool inspect_only;
/* Multi-GPU placement uses this to price per-layer KV storage. */
int placement_ctx_hint;
/* Number of independently allocated session graphs/caches to reserve. */
int placement_session_count_hint;
/* Server batch mode serializes execution and can share prefill scratch. */
bool share_session_prefill_workspace;
bool first_token_test;
@@ -371,6 +374,30 @@ int ds4_test_sample_logits(const float *logits, uint32_t n_vocab,
float temperature, int top_k,
float top_p, float min_p, uint64_t *rng,
float *prob_scratch);
int ds4_test_sampling_probabilities(const float *logits, uint32_t n_vocab,
float temperature, int top_k,
float top_p, float min_p, float *probs);
int ds4_test_speculative_sample(const float *target_logits,
const float *draft_logits,
uint32_t n_vocab,
float temperature,
int top_k,
float top_p,
float min_p,
uint64_t *rng,
float *target_probs,
float *draft_probs);
int ds4_test_speculative_delta_sample(const float *target_logits,
uint32_t n_vocab,
int draft_token,
float temperature,
int top_k,
float top_p,
float min_p,
uint64_t *rng,
float *target_probs);
int ds4_test_argmax_excluding_logits(const float *logits, uint32_t n_vocab,
int excluded_id);
uint64_t ds4_test_mixed_native_count(void);
#endif
int ds4_session_top_logprobs(ds4_session *s, ds4_token_score *out, int k);
@@ -403,6 +430,15 @@ int ds4_session_eval_speculative_argmax(ds4_session *s, int first_token,
int max_tokens, int eos_token,
int *accepted, int accepted_cap,
char *err, size_t errlen);
/* Evaluate one already-sampled target token and speculatively extend it.
* Positive-temperature DSpark normally commits greedily verified draft
* tokens; dspark_exact_sampling selects exact stochastic p/q acceptance. */
int ds4_session_eval_speculative(ds4_session *s, int first_token,
int max_tokens, int eos_token,
float temperature, int top_k,
float top_p, float min_p, uint64_t *rng,
int *accepted, int accepted_cap,
char *err, size_t errlen);
/* TP worker side of a mirrored speculative-verify block: run its half of the
* batch verify for KV side effects, then obey the leader's commit frame
* (keep, or roll back and replay). Only called from ds4_tp_worker_run. */