GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
Baekpica
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
Baekpica
Model Tree
deepseek-ai/DeepSeek-V4-Flash-0731
Input Modalities
Output Modalities
Supported Functionality
| Property | Value |
|---|---|
| Main-model parameters | 119,821,633,111 |
| Active parameters/token | 13,801,959,252 |
| Decoder layers | 43 |
| Routed experts/layer | 104 (from 256) |
| Experts selected/token | 6 |
| Stored floating dtype | BF16 Parameters; FP32 router-correction buffers |
The source repository's DSpark/MTP speculative-decoding auxiliary weights are
not included. The standard autoregressive causal model is preserved, and
num_nextn_predict_layers is set to 0.
Calibration used 1,024 sequences of 512 tokens (524,288 tokens total), sampled
and converted to the DeepSeek-V4 wire format from the private
Baekpica/K-EXAONE-236B-REAP-calibration-mix
dataset, revision 94e90afc261634a30384d3037d9021f1c997b286.
group_size=16).Implementation references:
SamsungSAILMontreal/ream, revision 84a3030716a0059589e9d10e2ea049e32b76cfa6Baekpica/solar-ream, revision d0201d1f484a521b4ec13a2e30a3ecb5a19a85b8Machine-readable build details are in ream_metadata.json; the final structural
audit is in checkpoint_audit.json, and hash-table remapping statistics are in
hash_router_audit.json. Per-layer direct-REAP distribution and route-coverage
diagnostics are in saliency_audit.json.
Healing used the private, audited
Baekpica/DeepSeek-V4-Flash-0731-REAM-Healing-Mix-2048
dataset. The four organic configs (seq1024, seq4096, seq8192, and
seq16384) are pairwise identity-disjoint. Two additional procedural configs,
needle8192 and needle16384, place exactly checkable records at front,
middle, and back positions in near-limit haystacks.
The merged adapter path contains 2,328 single-example optimization steps:
| Phase | Steps | Context used |
|---|---|---|
| Base instruction/code/multilingual healing | 1,024 | up to 1,024 |
| Identity-disjoint general healing | 1,024 | up to 2,048 |
| Selective long healing | 128 | 3,073–4,096 |
| Identity-disjoint reasoning/math healing | 128 | up to 2,048 |
| Balanced near-limit needle healing | 24 | 7,357–7,874 |
Rank-8, alpha-16 residuals were trained on 467 attention/shared-expert linear
modules, while the 43 native router matrices were updated at a much lower
learning rate. Routed 3-D expert tensors remained frozen. The final residuals
were merged into BF16 weights and the router matrices were saved directly.
healing_metadata.json records every phase report, dataset checksum, adapter
checksum, learning-rate provenance, and runtime_adapter_required=false.
needle16384 (14,737–15,726 tokens) is published as prepared training data but
was not used in this checkpoint; the final near-limit phase used only
needle8192.
Transformers 5.14.1 or newer is recommended.
python
from transformers import AutoModelForCausalLM, AutoTokenizerfrom transformers.models.deepseek_v4.modeling_deepseek_v4 import (DeepseekV4PreTrainedModel,)model_id = "Baekpica/DeepSeek-V4-Flash-0731-120B-REAM-104E-BF16"tokenizer = AutoTokenizer.from_pretrained(model_id)# Keep all learned Parameters in their stored BF16 dtype. The 40 non-Parameter# router-correction buffers remain FP32 for routing stability.DeepseekV4PreTrainedModel._keep_in_fp32_modules_strict = set()model = AutoModelForCausalLM.from_pretrained(model_id,dtype="bfloat16",device_map="auto",)
DeepSeek-V4 does not use a conventional Jinja chat template. The source
encoding/ utilities are included for prompt encoding and output parsing.
Expert merging changes model behavior, and short healing does not restore all
source behavior. A deterministic smoke prompt, 17 * 23, returned 289
instead of 391 after healing.
Two held-out eight-domain generation passes are included as
final_generation_smoke_8cases.json (512-token cap) and
final_generation_smoke_8cases_2048.json (2,048-token cap). After correcting
an erroneous math answer key in the first checker, the 512-token pass fully
satisfied 6/8 case rubrics (23/26 checks), while the 2,048-token pass satisfied
7/8 (24/26 checks). The Python LRU response improved from a truncated 512-token
answer to a complete 927-token response with EOS and all four code checks. The
7,440-token held-out retrieval prompt returned the exact requested JSON and all
three checks passed. Korean instruction formatting, structured JSON, English
model-systems explanation, and the science response also completed with EOS.
The quantitative-reasoning case is a material limitation. Its correct capacity
is 900 liters, but the model made an arithmetic error and settled on 600 liters,
then continued a repetitive self-correction loop to the full 2,048-token cap
without EOS (72.27% repeated 4-grams). The report preserves both the original
checker outputs and the corrected quality_check_version provenance. These
observations are diagnostics, not benchmark scores.
The procedural needle data and held-out generation checks do not establish general long-context quality. Reasoning delimiters, language selection, factuality, repetition, and task-specific accuracy should be evaluated on the intended workload. No claim is made that this checkpoint preserves the source model's published benchmark scores. The omitted DSpark/MTP auxiliary means it must be served autoregressively without the source model's MTP speculative configuration. The source model's MIT license and usage considerations apply.