Publish Fluid-1 Mini V21 R2 Step 180 Q6 DFlash
Browse filesValidated Q6 DFlash checkpoint. Existing license files and prior immutable revisions are preserved.
- config.json +20 -21
- model.safetensors +2 -2
- model.safetensors.index.json +1 -2
config.json
CHANGED
|
@@ -2,9 +2,24 @@
|
|
| 2 |
"architectures": [
|
| 3 |
"Qwen3_5ForConditionalGeneration"
|
| 4 |
],
|
| 5 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
"image_token_id": 98290,
|
| 7 |
-
"model_name": "
|
| 8 |
"model_type": "qwen3_5",
|
| 9 |
"pad_token_id": 98278,
|
| 10 |
"quantization": {
|
|
@@ -22,7 +37,7 @@
|
|
| 22 |
"attention_dropout": 0.0,
|
| 23 |
"attn_output_gate": true,
|
| 24 |
"bos_token_id": null,
|
| 25 |
-
"eos_token_id":
|
| 26 |
"full_attention_interval": 4,
|
| 27 |
"head_dim": 256,
|
| 28 |
"hidden_act": "silu",
|
|
@@ -69,7 +84,7 @@
|
|
| 69 |
"num_attention_heads": 8,
|
| 70 |
"num_hidden_layers": 24,
|
| 71 |
"num_key_value_heads": 2,
|
| 72 |
-
"pad_token_id":
|
| 73 |
"partial_rotary_factor": 0.25,
|
| 74 |
"rms_norm_eps": 1e-06,
|
| 75 |
"rope_parameters": {
|
|
@@ -93,21 +108,5 @@
|
|
| 93 |
"unsloth_version": "2026.4.6",
|
| 94 |
"video_token_id": 98291,
|
| 95 |
"vision_end_token_id": 98288,
|
| 96 |
-
"vision_start_token_id": 98287
|
| 97 |
-
"dflash_config": {
|
| 98 |
-
"block_size": 8,
|
| 99 |
-
"mask_token_id": 98278,
|
| 100 |
-
"target_layer_ids": [
|
| 101 |
-
2,
|
| 102 |
-
12,
|
| 103 |
-
21
|
| 104 |
-
],
|
| 105 |
-
"draft_vocab_size": 8192,
|
| 106 |
-
"num_layers": 5,
|
| 107 |
-
"quantization": {
|
| 108 |
-
"bits": 6,
|
| 109 |
-
"group_size": 64,
|
| 110 |
-
"mode": "affine"
|
| 111 |
-
}
|
| 112 |
-
}
|
| 113 |
}
|
|
|
|
| 2 |
"architectures": [
|
| 3 |
"Qwen3_5ForConditionalGeneration"
|
| 4 |
],
|
| 5 |
+
"dflash_config": {
|
| 6 |
+
"block_size": 8,
|
| 7 |
+
"draft_vocab_size": 8192,
|
| 8 |
+
"mask_token_id": 98278,
|
| 9 |
+
"num_layers": 5,
|
| 10 |
+
"quantization": {
|
| 11 |
+
"bits": 6,
|
| 12 |
+
"group_size": 64,
|
| 13 |
+
"mode": "affine"
|
| 14 |
+
},
|
| 15 |
+
"target_layer_ids": [
|
| 16 |
+
2,
|
| 17 |
+
12,
|
| 18 |
+
21
|
| 19 |
+
]
|
| 20 |
+
},
|
| 21 |
"image_token_id": 98290,
|
| 22 |
+
"model_name": "qwen35-2b-v21-r2-step180",
|
| 23 |
"model_type": "qwen3_5",
|
| 24 |
"pad_token_id": 98278,
|
| 25 |
"quantization": {
|
|
|
|
| 37 |
"attention_dropout": 0.0,
|
| 38 |
"attn_output_gate": true,
|
| 39 |
"bos_token_id": null,
|
| 40 |
+
"eos_token_id": 98278,
|
| 41 |
"full_attention_interval": 4,
|
| 42 |
"head_dim": 256,
|
| 43 |
"hidden_act": "silu",
|
|
|
|
| 84 |
"num_attention_heads": 8,
|
| 85 |
"num_hidden_layers": 24,
|
| 86 |
"num_key_value_heads": 2,
|
| 87 |
+
"pad_token_id": null,
|
| 88 |
"partial_rotary_factor": 0.25,
|
| 89 |
"rms_norm_eps": 1e-06,
|
| 90 |
"rope_parameters": {
|
|
|
|
| 108 |
"unsloth_version": "2026.4.6",
|
| 109 |
"video_token_id": 98291,
|
| 110 |
"vision_end_token_id": 98288,
|
| 111 |
+
"vision_start_token_id": 98287
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 112 |
}
|
model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e9604b175055d274238cfba11475b62269fbf9fc5132ec794ca8cc2ec18ec8ca
|
| 3 |
+
size 1517049312
|
model.safetensors.index.json
CHANGED
|
@@ -1,7 +1,6 @@
|
|
| 1 |
{
|
| 2 |
"metadata": {
|
| 3 |
-
"total_size": 1516941952
|
| 4 |
-
"total_parameters": 408672576
|
| 5 |
},
|
| 6 |
"weight_map": {
|
| 7 |
"language_model.dflash.d2t": "model.safetensors",
|
|
|
|
| 1 |
{
|
| 2 |
"metadata": {
|
| 3 |
+
"total_size": 1516941952
|
|
|
|
| 4 |
},
|
| 5 |
"weight_map": {
|
| 6 |
"language_model.dflash.d2t": "model.safetensors",
|